Lymphoproliferative response as an uveaLymphoproliferative response as an index ofcellular immunity...

British Journal of Ophthalmology, 1980, 64, 576-590

Lymphoproliferative response as an index of cellularimmunity in malignant melanoma of the uveaand its correlation with the histological featuresof the tumourM. S. NOOR SUNBA,1 A. H. S. RAHI, G. MORGAN, ANDE. J. HOLBOROW2From the 'Department of Pathology, Institute of Ophthalmology, University of London,and the 2Bone and Joint Disease Research Unit, The London Hospital

SUMMARY Cell-mediated immunity and immune complexes were studied in 22 cases of malignantmelanoma of the choroid and in 3 patients with iris melanoma. An attempt has been made for thefirst time to correlate the cell-mediated immunological changes with the histological features ofthe tumour and the extent of the disease. In most of the patients the lymphoblastic response totuberculin and phytohaemagglutinin was within normal range, indicating that there is no obvioussuppression of T lymphocyte-dependent immunity in intraocular melanoma. However, only abouthalf the patients showed an increased lymphoproliferative response to their own tumour antigensor to pooled allogeneic melanoma extracts. The response to the latter was invariably weaker,suggesting that although melanoma antigens may show cross-reactivity there exists an element ofindividual specificity. As the transformation index was generally low (i.e., -3) it is tempting to con-

clude that uveal melanomas are either poorly immunogenic or that the immune system is insuffi-ciently stimulated when the tumour is confined within the eye ball. The latter would seem more

likely. Although the lymphocyte transformation test appears to be unreliable in the early diagnosisof intraocular melanoma, it promises to be useful in the prognosis of this tumour. The host immunesystem appears to be better stimulated if the tumour is large and contains a detectable proportionof epithelioid cells. Histochemical and ultrastructural examination of the lymphocytes infiltratingthe tumour showed evidence of cytoplasmic activation, suggesting that these cells may be activelyengaged in tumour homoeostasis. With one exception Clq-binding immune complexes were notraised in the blood of the patients examined so that tumour-enhancement (through this mechanismat least) may not be an important feature in the natural history of intraocular melanoma.

Interest in the natural history of malignantmelanoma of the uvea and in the existence of a hostdefence mechanism has been stimulated by theobservation that uveal melanomas may remainlocalised for long periods-over 30 years in 1 case.1It is also known that malignant melanomas of theiris rarely metastasise," while some intraocularmelanomas may even undergo spontaneous regres-sion.45 In an attempt to explain this puzzlingbehaviour of some uveal melanomas several hypo-theses have been put forward, such as the effects ofthe host environment,6 defensive substances,7 bio-

Correspondence to Dr A. H. S. Rahi, Institute of Ophthal-mology, Judd Street, London WC1H 9QS.

chemical barriers,8 and specific immunemechanisms.9Although much of the antigenicity of tumour

extracts is attributable to components shared withnon-neoplastic cells, there are also antigens whichare peculiar to tumour cells in general and uvealmelanoma in particular.9 These tumour-associatedantigens are conveniently divided into 2 categories:those associated with the surface membrane andthose linked with the intracytoplasmic contents.The surface membrane antigens can initiate atransplantation rejection process, so that theseantigens have also been called tumour-specifictransplantation antigens. Both surface and intra-cytoplasmic tumour-associated antigens have been

576

on February 22, 2020 by guest. P

rotected by copyright.http://bjo.bm

j.com/

Br J O

phthalmol: first published as 10.1136/bjo.64.8.576 on 1 A

ugust 1980. Dow

nloaded from

http://bjo.bmj.com/

Lymphoproliferative response as an index of cellular immunity in malignant melanoma of the uvea

IMMUNOLOGICAL MECHANI SMS IN TUMOUR CONTROL

Normal Cell

Tumour Associated Sur(TASAl

_0000-0 Tumour Associated Cyt.0_Neoplastic Cell (TACA)

Chalone, O(-Fetoprotein -0

Immiinp Resnanse (toTASAKer Tumour Products

I -//I * Specific AntibodyI Activation ofI Suppressor Cells

Inhibition oflacrophages and |Lymphocytesi ~~~K-Cell( N

I I;

I I 'i

Tumour |CT.I Enhancement

'* Blocking Phenomer

Sensitized T-Lympho(

Immune Reaction

lK-Cell~jv'\; < Armed Macrophage

IN; b Cyti

Damaged Cell

Necrosis and

P Tumour Reduction

T Vasculitis-

non

Immuno-SuppressiveSerum Factors

0( -globulinProstaglandin E2

Fig. 1 Tumour-associatedsurface antigens (TASA)lead to both humoral andcellular immune responses.An antibody-mediated

irface Antigens reaction on the surface of aneoplastic cell may lead todirect cell damage either

toplasmic Antigens through the lytic activity ofthe complement or through aK cell and macrophage-related response. ActivatedT lymphocytes are cytotoxic

cyte to cancer cells. Tumournecrosis is accentuated bythe release of inflammatorylymphokines, some of whichare known to activate andarm macrophages to destroyneoplastic cells. The releaseof tumour-associatedcytoplasmic antigen (TACA)

Release of from damaged cells furtherwplasmic Antigens stimulates the immune system

( TACA and leads to the formationofimmune complexes which

ntibody Production either produce local inflam-matory reaction and tumournecrosis or lead to theblockade of T cell receptorsand tumour enhancement.

Immune-Complex This tumour enhancementFormation

may also result from severalother humoral factors and

,ntibody Production various tumour cellAtibody products.

TSurface Antigens (TASAI

described in malignant melanoma of the choroid,9 10and it would seem that, whereas there is a consider-able sharing of intracytoplasmic tumour antigensbetween one uveal melanoma and another, mostsurface antigens are peculiar to an individualtumour. It is possible, however, that melanomas ofthe uvea may resemble those of the skin in having2 additional types of surface antigens: one beingspecific to melanoma cells but common to a numberof individual tumours (histotype-specific) and theother shared with neoplastic and non-neoplastictissues in a number of species."-'3 Since antibodiesdo not penetrate intact cell membranes, the role ofcytoplasmic antigens in stimulating an effective hostdefence mechanism is questionable. Converselyantibodies to surface membrane antigens appear tobe cytotoxic to uveal melanoma cells in short-termtissue culture studies14 and are more likely to be

involved in host immunity. It is now accepted that,as with graft rejection, the control of tumourgrowth is mainly executed by T lymphocytes, whichdiffer from antibody in not requiring complementto exert cytotoxic activity and in being able topenetrate solid tissues (Fig. 1). Delayed cutaneoushypersensitivity to soluble extracts of melanomacells have been studied in patients with intraocularmalignant melanoma.'5 Although the majority ofthe patients tested showed a positive skin reactionto the test antigen, the reaction was negative in a

significant proportion of histologically confirmedpatients, and false positive results were obtained inabout 20% of patients with benign or non-

neoplastic intraocular lesions.Although the skin test appears to correlate well

with cellular immunity, particularly against infectiveorganisms, the expression of skin reactivity itself

and othf

M

577



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/

M. S. Noor Sunba, A. H. S. Rahi, G. Morgan, and E. J. Holborow

depends not only on the infiltrating T lymphocytebut also on the vasculature and the microanatomyof the test site and the capacity of an individual tomount a nonspecific inflammatory response follow-ing injury.'6 To overcome these in-vivo limitations

it has been suggested that in-vitro tests should becarried out. Manor and co-workers'7 thereforestudied direct (i.e., 1-stage) macrophage migrationinhibition tests in patients with uveal melanoma inwhich a crude extract obtained from 4 allogeneic(i.e., homologous) intraocular melanomas wereused as a source of antigen. This test seemed toprovide good discrimination between malignant andbenign pigmented tumours. However, its specificityas an indicator of cellular immunity is now beingquestioned because tissue extracts may either becytotoxic to leucocytes and cause nonspecificmigration inhibition or contain other substancessuch as antibodies which may alter the cell membraneof these cells and inhibit their movement.'8 It hasbeen suggested, therefore, that indirect (i.e., 2-stage)migration inhibition assays (in which the macro-phages are incubated with the supernate obtainedfrom a culture of patients' lymphocytes incubatedseparately with the test antigen) should providemore reliable results.'9

Cell-mediated immunity in malignant tumoursincluding those of the uvea has also been studiedby a macrophage electrophoretic mobility (MEM)test.20'2 The technique is based on the concept thatlymphocytes from patients with cancer are sensi-tised to a basic protein present in human brain andcancer cells. When incubated with this antigen, thesensitised lymphocytes release a macrophage-slowing factor which can be detected by its effect onthe electrophoretic mobility of guinea-pig macro-

phages. This technique differs from the conventionalmacrophage migration inhibition test in that in theMEM test it is the passive movement rather thanthe active migration which is grossly affected.Although the validity of this in-vitro blood test hasbeen confirmed,2223 several other workers, includingourselves, have been unable to do so and foundthat technical limitations which make standardisa-tion difficult render the test of little value.2'24-26

Malignant melanoma of the uvea can posediagnostic problems despite the development ofsuch sophisticated techniques as fluorescein angio-graphy, the 32P test, and ultrasonography. A mis-taken diagnosis of malignancy is more common insmall tumours of the uvea,2' and the incidence ofunnecessary enucleation may be as high as 20%.2830Because of this there is a considerable incentive todevelop a reliable technique for the preoperativediagnosis of malignant melanoma.The serological diagnosis of an intraocular

malignant melanoma appears to be difficult becauseof the presence of ostensible antimelanoma anti-bodies in patients with non-neoplastic eyedisease,1213 and it seems unlikely that antibodiesalone are responsible for tumour homoeostasis. Wedecided, therefore, to investigate other parametersof host immunity to tumours. Since skin tests,leucocyte migration inhibition tests, and macrophageelectrophoretic mobility tests have been carried outwith variable success in the diagnosis of intraocularmalignancy, and it has been reported that thelymphocyte transformation test using radioactivethymidine is easy to quantitate and correlates wellwith the extent of the disease in relation to malignantmelanoma of the skin,3' we have attempted toevaluate the usefulness of the latter test as animmunodiagnostic procedure.

Methods and materials

Fresh tumour cells obtained from enucleated eyeswere homogenised in sterile phosphate-bufferedsaline (pH 7 4) with a pestle and mortar and thentransferred to an electrical blender for furtherhomogenisation. The suspension was centrifuged at2000g for 10 minutes to remove cellular debris,nuclei, and melanin pigment. It was then dialysedagainst phosphate-buffered saline (pH 7-4) for 24hours at 4°C and stored in liquid nitrogen untilused. Antigenic extracts were also prepared usinghypertonic potassium chloride solution as originallydescribed by Meltzer et al.32 The cell suspensionobtained from minced tumour material was sus-pended in 20 ml of 3 M KC1 (pH 7 4) and left at4°C for 24 hours with frequent shaking. The sus-pension was centrifuged at 40 000 g for 60 minutesand the supernate was dialysed for 16 hours againstphosphate-buffered saline (pH 7-4). The dialysedmaterial was centrifuged again at 40000 g for 15minutes. The supernatant was concentrated byusing Lyphogel and sterilised by passing the solutionthrough a Millipore filter (0-22 sum). The proteinconcentration was estimated by a modified Lowrymethod.33

LYMPHOCYTE CULTUREBlood was collected before enucleation in preserva-tive-free heparin (20 units per ml) and lymphocyteswere separated by a density-gradient techniqueusing a Ficoll-Hypaque preparation (Pharmacia).After the viability of lymphocytes had been testedby the trypan blue technique they were cultured inmedium 199, each millilitre of which contained 200units of penicillin and 100 ,ug of streptomycin. Themedium was enriched immediately before use by

578



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/


the addition of the patient's own serum (1:10) andfreshly prepared glutamine (2 mmol/ml). The testwas carried out in Sterilin tubes in a gaseous phaseof 5% carbon dioxide in air; 106 lymphocytes sus-pended in 1 ml of the enriched nutrient mediumwere put in each tube. In an initial study variousconcentrations of the crude preparation as well aspotassium chloride extracted antigens were used toobtain a dose-response curve. Suitable amounts ofboth autochthonous (i.e., autologous) and pooledallogeneic (i.e., homologous) antigen preparationswere added separately to the various culture tubes24 hours after the initial incubation of the lympho-cytes, during which time the suppressor lympho-cytes appear to become less effective and do notgrossly interfere with the lymphoproliferative res-ponse of antigen-sensitive lymphocytes (Brostoff,personal communication). In order to study thestate of cellular immunity in general, the lympho-cytes from patients with malignant melanoma ofthe uvea were also cultured with various concentra-tions of specific and non-specific mitogens such aspurified protein derivative (PPD) and phytohaemag-glutinin (PHA). Controls to exclude a possiblemitogenic effect of normal human choroid were setup and the various tests were carried out in duplicate.

In order to study the specificity of this testmelanoma antigens were also added to the lympho-cyte from healthy individuals, from patients withnonmalignant eye diseases, including disciformmacular degeneration, sympathetic ophthalmitis,melanocytoma (magnocellular naevus) of the ciliarybody, and naevus of the iris, and also from a caseof carcinoma of lung with metastases in the choroid.2 ,uCi of tritiated thymidine (specific activity 20 Ci/mmol) were added to each control and the experi-mental tubes on the 6th day of the experiment, andincubated for a further period of 16 hours. Thelymphocytes were then harvested by centrifugation;the supernatant was deep frozen and studied laterfor evidence of lymphokine production. The lym-phocytes were resuspended in phosphate-bufferedsaline and passed through a specially designedfunnel containing fibreglass filter paper. They werewashed with 5 ml of 10% trichloroacetic acid and20 ml methanol. The filter paper was transferred toscintillation vials and dried in an oven at 37°C for3 hours; 10 ml of toluene-based scintillation fluidwas added to each vial, and the lymphoproliferativeresponse was measured in terms of 3H thymidineuptake (for DNA synthesis) in a Packard tricarbliquid-scintillation counter. The disintegration wasexpressed in counts per minute. To obtain thetransformation index the count in the vial con-taining antigen-treated cells was divided by theradioactivity in that containing unstimulated cells.

INDIRECT MACROPHAGE MIGRATIONINHIBITION TESTMacrophages were obtained from adult maleHartley guinea-pigs 6 days after the injection of25 ml of sterile liquid paraffin into the peritoneum.Heparinised Hanks's fluid was injected into theperitoneal cavity, and the macrophage-rich suspen-sion was aspirated with a 20 ml syringe attached toa perforated polypropylene tube, care being takennot to traumatise the tissues. The cell suspensionwas passed through a separating funnel, spun at250 g for 10 minutes, and resuspended in medium199. The concentration was -djtisted to 8x 106 cellsper millilitre of the medium, v.hich was enrichedwith normal guinea-pig serum. rhe indirect (i.e.,2-stage) inhibition studies were carried out inspecially designed Sterilin chambers. During thestudy it was found that Plasticine used to seal thecapillary tubes produced nonspecific inhibition ofmacrophages, and therefore a small amount ofsterile silicon gel was introduced into the capillarytube to separate the macrophage column from thePlasticine. The cells were incubated in the super-natant obtained from the unstimulated culture, andthe area of migration was compared with those insupernatants from antigen-treated cells. Supernatantfrom antigen-treated lymphocyte cultures fromhealthy individuals was also used in a separatestudy to exclude any nonspecific inhibitory effect ofmelanoma extract.

Migration inhibition of more than 25% giving anindex of less than 0-75, was regarded as abnormalfor the present study.

IMMUNE-COMPLEX ESTIMATIONIt has been shown that a tumour may sometimesshow a marked increase in size in the presence ofan apparent immune response. This is known astumour enhancement, and it is now considered tobe due to blocking of antigen-specific receptors onT lymphocytes by free tumour antigens or antigen-antibody complexes.34 It is possible, therefore, thatthe in-vitro lymphoproliferative responses to tumourantigens may also be modified by the presence ofantigen-antibody aggregates in the patients' serum.35Sera from melanoma patients was investigated,therefore, for the presence of soluble immune-complexes using an 1251-Clq binding assay. Clqwas isolated from normal human serum and radio-iodinated by the lactoperoxidase method. Fixedamounts of radiolabelled Clq were added to thetest serum, and the complexes were precipitated with3% (w/v) polyethylene glycol (PEG). The mixturewas centrifuged at 1500 g for 20 minutes at 4°C, thesupernatant was discarded, and the radioactivity ofthe precipitate was measured. Results were expres-

579



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/


sed as a percentage of '25I-Clq precipitate relativeto the radioactivity in the control tube in whichnormal serum containing the same amount ofradiolabelled Clq is precipitated with 20% tri-chloroacetic acid (TCA). Blood from 50 healthyadults and 52 cases with various eye diseases wasalso tested for immune complexes, and the dataobtained were used for comparison and statisticalanalysis. In the present study only those samplesthat showed 1251-Clq binding of 12% or over wereregarded as abnormal and were considered tocontain raised levels of immune complexes.

HISTOLOGICAL EXAMINATIONAfter removal of part of the tumour for immuno-logical studies the enucleated eyes were fixed inbuffered formol saline, and paraffin sections werestained with haematoxylin and eosin and methylgreen and pyronine Y for histological examination.The sections were bleached in cases where the pig-ment content made it difficult to assess the cell type.The largest diameter of the tumour in contact withthe sclera was measured in millimetres. The sectionswere also examined for evidence of necrosis,lymphocytic infiltration, immunoblastic transfor-mation (i.e., pyroninophilia), and extraocular exten-sion. A portion of the tumour was first fixed ingluteraldehyde and then in osmium tetroxide andblocked in Araldite for electron microscopicalexamination. The sections were stained with leadcitrate and uranyl acetate and examined with aJOEL IOOC transmission electron microscope.Statistical analysis of the immunological investiga-tions was carried out by Fisher's exact test.

Results

A transformation index of 2 or more was regardedas evidence of sensitisation to tumour antigens. Thelymphocyte transformation test using autochthanous(i.e., autologus) tumour extract was performed inonly 19 of the 22 cases of malignant melanoma ofthe choroid because enucleation was not undertakenin the other 3 cases. An increased lymphoprolifera-tive response was seen in 10 cases (Table 1). Theresponse to allogeneic tumour antigen, however,was weaker than that incurred by autologoustumour extract. In general, the response to KC1extracted antigen was marginally stronger than thatto the crude antigen preparation. There was also awide variation in the response to PPD and PHA,but with the exception of 5 cases (see Tables 1 and3) the transformation index was generally withinthe normal range (that is, 312 for PHA).

Generally it was found that patients who showed

marked reactivity towards PPD and PHA alsoshowed a response to tumour antigens. It can beargued that a generally low uptake of radioactivityin our experiment was due to the fact that 3H-thymidine of a high specific activity (20 Ci/mmol)was used, since this might have damaged some ofthe lymphocytes. We have used, however, the sameisotope with a low specific activity, i.e., 2 Ci/mmolin 6 cases without any marked changes in thelymphoproliferative response. The lymphoprolifera-tive response to normal human choroidal extractwas studied in only 9 patients, and the transforma-tion index in these cases was always <2. This is inkeeping with the previous report in which it wasshown that antimelanoma antibodies do not cross-react with normal choroid.36The lymphocyte transformation test was carried

out in only 3 cases of malignant melanoma of theiris; autochthonous (autologous) tumour antigenwas used in 2 patients, and a high thymidine uptakewas present in both cases (Table 2). The reactivitytowards allogeneic tumour was weaker, and in only1 case was there an appreciable increase in theDNA synthesis. The response to PHA was withinthe normal range in 2 of the 3 cases examined.The lymphoproliferative response to allogeneic

melanoma extract was minimal or negligible inhealthy controls as well as in patients with clinicallybenign lesions, secondary carcinoma of the choroid,and in non-neoplastic eye disease. The response toPHA in most of these cases was normal (Table 3).The statistical analysis of the lymphoproliferativeresponse and its correlation with various histologicalparameters are summarised in Table 4.

In cases where the lymphocyte transformationtest was positive (Table 1) it was found that thetumour was generally large (P<0001), the largestdiameter in contact with the sclera being 12 mm orover, though there was 1 exception in which atumour with a diameter of 15 mm showed a trans-formation index with autochthonous tumour extractof less than 2. In the case of iris melanoma a positiveresponse was associated with a diffuse infiltratingtumour.

Extraocular extension was evident in 6 cases ofmalignant melanoma of the choroid and in 1 patientwith iris melanoma. The lymphocyte transformationtest was positive in all of them (P<0-001).

In only 1 case was the tumour of epithelioid celltype, 8 cases were of mixed variety, and the restwere composed almost exclusively of spindle cells.The lymphoproliferative response was increasedonly in patients with a mixed-cell type of melanoma(P<0001).

Pigmentation was heavy in 4 cases, moderate inanother 6, and light in the rest. There was no corre-

580



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/


+ # * * * * * * * * * * * * * * * * #

eq 0 0 vo0 - 0 e 0oNo N N N N N - xMC 0 0It e t WI WI ItO NItdn V't - --t -

* * * * Nt C .0* * * * * * * * * -_* * * *

0 0o0o0o r 00 00 0 0 00o 00 00 0 0* NC M M 4l 44 el 000 - - 0o 0 -00

0 0 M It " 0 en 0 0Co 00o 00 0-c, N N a, 00xces4 CN " C^ N 4CN en̂ o O O O O - -, O O O # # #

++ +++ + + ++ d I- I 1flll+ 1

+ + ++ + + +-+ ++ + + 1- +- + + + + +

II+ I + + I + + III

++ + ++++ I I-+I +

±

>0000Z-4

0z

>1Oa000 000

o oZ z

a 4)

00 00000t 00 0 0

x x. x ° x, ,xcccccccc

EEE E E E E E E E E E E E E E E E EE E E E E E E E E E E E E E E E E EeninenNC4 It eq It en in~~~t_ C:yxo in t_

eJ~~ ~ ~ ~ tQo t- ;;NQ--E-- E -- - - - - - - - - --

as'(U Y Y Y Y Q

.0.0._.0-.0---

+ #

etts

*c00 E

eit

0t

( '42- 80 0 0

.M 0

+-

+

0

-0

.0

a.-

a.

0t

.-0tU)

0,G

0,u0,m<

a U'-

*S 00 a

.000,0. o a

'4 x~(t -0

'-g.

581

4;

0

0

z

a-

-04.e00.-

-0

-0

U)

-0

-0

0d

>0ba0-5.0

0z

(U

0C0._2:lz

.0004>d'U1

C 0X

A 4,._s 1: .':3 Roso4

z

.2- M a0es L'

Q_ v.e

I

1;D3

D



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/


Table 3 Lymphoproliferative response to melanomaantigens in the control group

Cellular immunity

Transformation index

No. Sex Nature of lesion Allogeneic(homologous)tumour PHAextract

1 Male Disciform macular 0-96 50degeneration

2 Female Sympathetic ophthalmitis 1-0 34

3 Male Lepromatous iritis 1-0 27

4 Male *Melanocytoma of the 093 20CB (magnocellularnaevus)

5 Male Iris naevus 0-8 13

6 Male *Lung carcinoma with 0 9 34choroidal metastases

7 Male *tMetastatic adenocar- 0-84 4-7cinoma (choroidal)

8 Male Healthy control 1-1 28

9 Male Healthy control 1 39



12 Male Healthy control 1-2 40

*Histologically confirmed. tA state of generalised immunosuppres-sion cannot be ruled out. CB = ciliary body.

lation between the degree of pigmentation and thelymphoproliferative response.

Distinct apoptosis and coagulative necrosis wereevident in 6 cases and absent in the remaining 15cases. The tumour-specific immune response ap-peared to correlate well with the extent of necrosis(P<O.OO1).Lymphocytic infiltration (Fig. 2) was extensive

in 2 cases, prominent in 8 cases, and occasional in2 cases. The remaining 9 cases did not show anyround cell infiltration in the sections examined.Methyl green pyronine Y staining showed thepresence of occasional immunoblasts among theinfiltrating lymphocytes, thus providing directevidence of in-vivo activation of these cells. Electronmicroscopic examination complemented this findingand revealed the presence of lymphocytes with welldeveloped rough endoplasmic reticulum and promi-nent mitochondria (Fig. 3).

Mitotic activity was markedly increased in 3cases, all of which had a positive lymphoprolifera-tive response. A moderate to low activity was seenin 7 patients, and the test was positive in only 2patients. The mitotic activity does not appear tocorrelate with the degree of blastic response.

1251-Clq binding assay was carried out in 22cases of malignant melanoma of the choroid and 3cases of iris melanoma and the results were com-

Table 4 Tumour-specific immunity in malignant melanoma of the uvea: analysis of significance

Characteristics No. tested No. positive No. negative Significance

Control (C) 12 0 12

<lOmm(DI) 7 0 7 DI vsCNSDiameter D2 vs C P<O0001

>10 mm (D2) 14 12 2 D2 vs Dl P<OOO0

Spindle (TS) 12 3 9 TS vs C NS

Cell type TM vs C P<O001Mixed (TM) 9 9 0 TM vs TS P<O005

Present (NI) 6 6 0 NI vs C P<O-O01Necrosis NO vs C NS

Absent (NO) 15 5 10 Nl vs NO P<0-05

Present(LI) 13 11 2 Ll vsCP<OOOILymphocytic LO vs C NSinfiltration

Absent (LO) 8 1 7 LI vs LO P<O OI

Present (Ml) 10 5 5 Ml vs C P<O05Mitosis MO vs C P<O05

Absent (MO) 11 7 4 Ml vs MONS

Present (Xl) 7 7 0 Xl vs C P<O-OOIExtraocular X sCNAbsente(XO)s14o5n9 XO vs C NS

Absent (XO) 14 5 9 Xl vs XO P<cO05

582



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/

Fig. 2 Section ofa ma!ignant melanoma of the choroid showing lymphocytic infiltration. A: Spindle cellmelanoma. (H and E, x 185). B: Mixed cell melanoma. (H and E, x 590).

Fig. 3 Transmission electron micrograph showing a melanoma cell (M) with an intranuclear cytoplasmic inclusion.The adjacent cell is a lymphocyte (x 10 670). The inset shows an activated lymphocyte with a cleaved nucleus andabundant cytoplasm containing ribosomes ( x 6400).



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/


Table 5 Immune complexes in the bloodfrom patientswith malignant melanoma of the uvea (125! Clqbinding assay)

No. No. positiveNature of the lesion tested (binding >12%)

Healthy adults 50

Degenerative eye disease 50Control

Melanocytoma ciliary body 1 1

Benign iris naevus 1

Malignant +ve LTT 10 1melanomachoroid -ve LTT 12

Malignant +ve LTT 2melanomairis -ve LTT 1

LTT = Lymphocyte transformation test (see Tables I and 2).

pared with those in the control group. With 1

exception the serum Clq binding activity in mela-noma patients was less than 12%, and thereforethe levels were regarded as lying within the normalrange (Table 5).The indirect macrophage migration inhibition

test was carried out in only 4 cases of malignantmelanoma of the choroid and in 1 patient with irismelanoma. The test was positive in 1 patient withchoroidal melanoma and in the patient with an irismelanoma (see Tables 1 and 2).

Discussion

Both humoral and cellular immunity have beendemonstrated in a variety of tumours in man,including malignant melanoma of the skin. It isgenerally considered, however, that cell-mediatedimmunity is primarily responsible for tumourrejection in experimental animals, and severalreports indicate that a similar mechanism may

operate in human neoplasms.37The antigens that elicit this type of reaction appear

to be located on the cell membrane and behave as

tumour-specific transplantation-like antigens. Sincein our experiments, apart from 3 MKC1 solubilisedmembrane proteins, we also used homogenisedtumour extract, it can be argued that the stimulatoryantigens were of both membranous and cytoplasmicorigin. It is possible, therefore, that the respondinglymphocytes in these 2 experiments belong toseparate subsets.As humoral immunity has already been demon-

strated in malignant melanoma of the choroid938the finding of T cell reactivity to tumour extractraises the question of their respective roles in thehost's defence mechanism against such tumours.

Although it is generally agreed that T lymphocytemediated immunity is of more significance intumour control, it is not certain whether the lym-phocytes which show increased DNA synthesis (asseen in the present study) represent cytotoxic Tcells or helper T cells which co-operate with Blymphocytes in the production of tumour-specificantibodies.The delayed hypersensitivity skin test remains

one of the simplest means currently available forthe clinical assessment of the status of cellularimmunity in man. Melanoma-associated antigensfrom allogeneic choroidal tumour extract havebeen used for the diagnosis of intraocular mela-noma.'5 Whereas 90% of the patients with histo-logically confirmed intraocular melanoma gave apositive skin response, about 20% of patients withnonmalignant diseases also showed a positivereactivity towards melanoma extract. The authorsargue that the false positive results may have beendue to contamination of the antigenic preparationswith normal tissue constituents against which thepatients in the control group had been sensitised,but this explanation is equally applicable to thosecases which the authors regarded as true positives.

It is now accepted that the skin test as a correlateof cellular immunity has several limitations. Undercertain circumstances the delayed cutaneous hyper-sensitivity may actually be depressed at a time whenthere is an intense cell-mediated reaction to thesame antigen elsewhere in the body. It is known,for example, that in experimental allergic orchitisthere may be a depression in skin test reactivity at atime when the test is diffusely infiltrated by lympho-cytes. This depression in skin reactivity is evidenteven when such non-cross-reacting antigen astuberculin is used.39

It is now well documented that a positive delayedskin test is dependent not only on a specific immuno-logical reaction but also on the patient's ability tomount a nonspecific inflammatory response.16 Theinterpretation of abnormal results, therefore, maysometimes be difficult. Furthermore, to determinewhether or not the patient has T cell mediatedimmunity to the test antigen, the individual shoulddevelop a skin reaction consisting of at least 5 mmor more of induration at the test site 48 hours afterthe intradermal injection of the antigen (as is thecase in Char's study on ocular melanoma). Itshould be noted, however, that such a reaction,especially if it is not strong, may not always repre-sent a cell-mediated hypersensitivity because itcould very well be a remnant of a previous Arthus-type reaction. In addition, the induration which isthe hallmark of delayed cutaneous allergy is in factdependent on vascular leakage and deposition of

584



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/


fibrin,40 which may be defective in some individualsand in those on anticoagulant therapy. Skin testsare often difficult to quantitate, and often it is noteasy to produce an absolutely sterile tumourextract for injection in man, especially with respectto possible contamination with hepatitis B virus.Furthermore, it may be impossible to exclude a

carcinogenic effect of tumour extracts when injectedinto healthy human volunteers used as controls.The alternatives to skin testing are macrophage

and leucocyte migration inhibition and lymphocytetransformation tests. As mentioned earlier, positiveinhibition tests have been reported in choroidalmalignant melanoma, indicating that a degree ofcellular immunity is present in these patients. Themacrophage migration inhibition test (MMT) inwhich guinea-pig peritoneal macrophages are usedmay be subject to variation either due to nonspecificreaction of the proteins in tumour extract with themacrophages directly or through the activation ofxenogeneic (heterologous) lymphocytes which oftencontaminate such preparations.

In addition it is well known that some antigensmay produce 'toxic' migration inhibition by inhi-biting both sensitised and nonsensitised cells.41 Ithas also been shown that migration inhibition testsare most successful when particulate antigens areused. Soluble antigens may fail to cause inhibitionof migration even when other tests give a positivecell-mediated reaction.4" Negative results shouldtherefore be treated with caution. Migration inhi-bition studies involving choroidal melanomasshould be assessed against this background.The lymphocyte transformation test (LTT)

appears to have an enormous range of application.Whereas the skin test assays predominantly theeffector limb of delayed hypersensitivity, which ismediated through a subset of T lymphocytes knownas (TDH), the LTT can detect the activity not onlyof these cells but also of the helper T cells (TH),which modulate the activity of the antibody pro-ducing B lymphocytes. Suppressor T lymphocytes(TS) are also activated during an in-vitro lympho-proliferative response.Although there are a variety of in-vitro tests to

evaluate cell-mediated immunity, the LTT providesa simple and relatively more reproducible and easilyquantifiable method. But this technique also has itsown limitations. A nonspecific lymphoproliferativeresponse can be induced by the constituents of theculture media as well as the contents of variousxenogeneic (heterologous) and allogeneic (homolo-gous) serum products; excessively vigorous handlingof lymphocytes may also stimulate T cells. We havetherefore used only autologous serum in theseexperiments. Since the uptake of radioactive thymi-

dine increases with the duration of incubationbecause of exponential cell growth, proper statisticalassessment is difficult unless data are normalised,i.e., expressed in such a way that a normal distri-bution is obtained again. Although the techniqueof expressing the result as transformation indicesrather than absolute counts normalises the data tosome extent, it is assumed that the degree of reac-tivity in the antigen-treated and nontreated culturetubes will be affected by the various variables inthe same way and will remain proportional. Unpre-ventable preincubation stimulation of the lympho-cytes may, however, at the same time elevate theresponse in the control tube and interfere with thereactivity of lymphocytes with the test antigen inthe experimental tube. For example, the patientmay be allergic to antibiotics used in the cultureserum. Fortunately, however, there was no reasonto suspect the latter complication in any of thecases in the present study.

Since the lymphoproliferative responses caninvolve both T and B lymphocytes, the LTT wouldappear to reflect a wider spectrum of host immunitythan either the skin test or the migration inhibitiontest, these tests depending almost entirely on thereactivity of T lymphocytes.The lymphoproliferative response to autochtho-

nous and allogeneic extract of malignant melanomaof the choroid were studied in 22 cases of choroidalmalignant melanoma. An increased DNA synthesiswas demonstrable in 12 of the 21 histologicallyconfirmed cases; in the remaining 4 cases, in whichthe diagnosis was based on clinical evidence alone,the lymphoproliferative response was negativewhen using allogeneic tumour extract.A positive lymphoblastic response to autoch-

thonous tumour extract was also associated withsome reactivity towards allogeneic tumour extracts,but the thymidine uptake was comparatively lowerin the latter series. This suggests that the responsewas due to antigens which, although showingfeatures of cross-reactivity, have some element ofindividual specificity. Whereas it is possible thatthe response to allogeneic tumour extract was dueto histoincompatibility, it would seem unlikely,because the thymidine uptake following treatmentwith allogeneic tumour extract was never higherthan uptake by the lymphocytes when challengedwith autochthonous (autologous) extract.The lymphoblastic response to phytohaemagglu-

tinin (PHA) was normal except in 5 cases in whichthe response to autochthonous or allogeneic tumourextract was not only poor but showed evidence of adegree of inhibition rather than proliferation. It istempting to speculate that this was due to thepresence of specific inhibitors of mitosis (i.e.,

585



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/


chalones42) in the tumour extract, as similar com-pounds are known to be present in malignantplasmacytoma.43 It has also been suggested thatautologous serum from cancer patients may have aninhibitory effect on in-vitro lymphoproliferativeresponses.44 While this could account for the rela-tively low transformation index in some of our

patients,45 this nonspecific inhibitory influencewhich is either due to ac-macroglobulin (Kamrin-Mowbray factor) or to nonspecific binding of antigento other plasma constituents is usually seen at a

serum concentration much higher than that usedin the present study.46 It has not been possible,however, to exclude the possible inhibitory role ofoc-fetoprotein in these cases.47 Furthermore, the in-vitro immunosuppressive effect of glass-adherentmononuclear cells also could not be ruled out48(Fig. 1), though precautions were taken to minimisethe effect of suppressor T cells. In a pilot study itwas found that serum from melanoma patientsshowing a poor response to PHA could also inhibitthe lymphoproliferative response in normal indi-viduals. This phenomenon is being investigated indetail. Since it has been suggested that the presenceof free antigens or antigen-antibody complexes inpatients serum may modify the antigen-specificlymphoproliferative response in tissue culture,attempts were made to detect the presence of suchcomplexes in the serum of the patients in whomlymphoproliferative responses were studied, as

shown in Table 4.The level of circulating immune-complex was

within the normal range except in 1 patient in whomthe tumour was markedly necrotic. Although raisedlevels of immune-complexes have been reported inretinoblastoma,49 which are invariably necrotic,Dernouchamps and associates50 using a differenttechnique found that only 2 of their 7 cases ofmalignant melanoma of the choroid had any

detectable amount of Clq binding aggregates in theblood. It would appear, therefore, that in malignantmelanoma of the choroid at least an immune-complex-induced tumour enhancement may notbe an important feature of its natural history. Thereason we used autologous serum instead of fetalcalf serum throughout this study was that the latteris known to produce a toxic effect after heat inacti-vation and may contain a substance which willproduce nonspecific stimulation of the lympho-cytes. Either effect could obscure any difference inthe radioactive uptake counts between the controland the experimental tubes.51 Serum from healthyvolunteers can be used as a substitute for autologousserum in order to exclude the presence of inhibitors,but this again is not free from problems, because theserum has to be fresh,51 and, furthermore, antibodies

in the allogeneic serum may react with blood groupantigens on the surface of the patients lymphocytesand modify their response to tumour antigens. Thislimitation would be reduced, however, if the patientsstudied belonged to blood group 0 and thereforelacked group A or B antigens on the cell surface.52Recent studies suggest, however, that immuno-suppressive substance present in cancer-bearinganimals have a high affinity for lymphocytes, andtherefore substitution of autologous serum isunlikely to alter the lymphoproliferative response.53It is of interest in this context that the antigen-specific lymphoproliferative response in cutaneousmelanomas is similarly of low key possibly becauseof weak immunogenicity of melanoma antigens.5455An indirect macrophage migration inhibition test

was carried out in only 5 cases of malignant mela-noma of the uvea. The test was positive in 2 casesin which the lymphocyte transformation index wasalso high.The migration inhibition test appeared to corre-

late with an in-vitro lymphoblastic response exceptin 1 patient in whom there was no evidence oflymphokine production in spite of a positivelymphocyte transformation test. Our findings donot compare with those of Manor and co-workers,"7who were able to demonstrate a macrophage migra-tion inhibition in 7 of their 8 cases of choroidalmelanoma. This may be due to the fact that theyused a direct (i.e., 1-stage) migration inhibition test,which is more likely to be affected by nonspecificfactors.The lymphocytes obtained from healthy donors

as well as patients with metastatic carcinoma,benign intraocular naevus, and non-neoplastic eyediseases did not show any appreciable reactivity topooled allogeneic melanoma extract, which furthersupports the conclusion that the in-vitro nuclearactivity seen in melanoma patients was tumour-specific and not due to stimulation by transplanta-tion antigens.The enhanced in-vitro activity of lymphocytes in

melanoma patients correlates well with the extentof the disease, being markedly increased in everycase which showed evidence of extraocular extensioncompared with only 5 of the 14 patients in whomthe tumour was localised to the uvea. It is knownthat the intraocular tissues are not provided withlymphatic channels, so that the afferent limb of theimmune response may be prolonged with respect toantigens originating within the eyeball. This maybe partly responsible for the immunological privi-lege enjoyed by the eye and cause delayed sensitisa-tion to intraocular malignancy, in which case theprivilege is a disadvantage to the individual. Thegenerally low transformation index in our patients

586



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/


Fig. 4 Scanning electronmicrograph ofa malignantmelanoma of the iris (x 516).The arrangement of tumourcells is such that duringcontraction of the iris some ofthem may detach into theanterior chamber, where theymay either become engulfed bythe trabecular endothelium orenter the Schlemm's canal toreach general circulation.

should therefore be seen in the light of this back-ground and not compared with tumours arising inother parts of the body, where a low transformationindex is usually associated with generalised metas-tases.56 57The cell-mediated immune response in uveal

melanoma appears also to be dependent on thesize of the tumour and hence on the amount ofantigens available. The lymphocyte transformationtest was positive only in those cases in which thelargest tumour diameter in contact with the sclerawas 12 mm or more, while in all the negative casesthe tumour was 10 mm or less in diameter apartfrom 2 cases where it was 11 mm and 15 mm res-pectively. It may also be of significance that inpatients with a negative lymphocyte transformationtest the tumour consisted of compact spindle cells,while in all the cases with a positive lymphocytetransformation test the tumour consisted of avarying mixture of spindle and epithelioid cells.Since epithelioid cells appear less cohesive, theymay more easily become detached and lead notonly to early metastasis5859 but also to stimulationof the immune system. It is possible that thisexplanation is not applicable to melanomas of theiris, where tumour cells of both spindle and epithe-lioid variety or their antigenic components may bereleased into the anterior chamber during normalcontraction and relaxation of the iris (Fig. 4) andpass into the canal of Schlemm either directly orafter being taken up by the resident population ofphagocytic cells in the trabecular meshwork, whichinclude both histiocytes as well as endothelialcells.60The growth and metastatic property of a malig-

nant tumour is greatly affected by the immuno-

genicity of the tumour and the immunologicalstatus of the host.6' Furthermore, it is known thatantigens become more immunogenic after beingprocessed by macrophages.62 It is conceivabletherefore, that antigenic substances released froma tumour of the iris would have the advantage ofhaving a head-start. This may be the reason why 2of the 3 cases of malignant melanoma of the irishad possible evidence of sensitisation to tumour-associated antigens as compared to malignantmelanoma of the choroid, in which only 45% ofthe patients were sensitised. Whether or not thispossible early sensitisation is responsible, at leastin part, for the slow growth and better prognosis inmelanomas arising from the iris remains to beproved.

It is possible that infiltration of the tumour bylymphocytes and plasma cells represents an immuneresponse aimed at eradication of a malignant lesion,such infiltration having been claimed to improvethe prognosis in a variety of tumours, includingthose arising in the breast.63 More recently Tritsch64produced evidence that an intense inflammatoryreaction at the periphery of cutaneous melanomasreflects a favourable prognosis.

It is significant that all 10 cases giving a positivelymphocyte transformation test showed varyingdegrees of lymphocytic infiltration in and aroundthe tumour, though it is not known whether theinfiltration simply reflected a state of hypersensi-tivity or represented a favourable prognostic sign.These patients will have to be followed up forseveral years before any conclusion can be drawn.Histochemical and ultrastructural changes in theinfiltrating lymphocytes are indicative of an in-vivoimmune reaction, and fewer morphological changes

587



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/


have been reported in cutaneous melanomas.6566We are studying the nature of the lymphocytesbecause it would be interesting to know whetherthese cells are cytotoxic lymphocytes engaged intumour homoeostasis or suppressor T cells that areinvolved in tumour enhancement.67-9

Contrary to an earlier belief, malignant tumoursare not truly autonomous, that is, capable of growingrelentlessly. Homoeostasis in cancer is now a wellrecognised entity, and a spontaneous regression ofcertain tumours such as choroidal melanoma isonly one aspect of this important phenomenon. Itwould appear that a tumour is a heterogeneous anduneven collection of cells, some of which areproliferating (i.e., taking part in the mitotic cycle)and some of which are not, and which is subject todepletion by migration and cell death.70 Necrosis isthus not uncommon in choroidal melanoma, anddoes not always appear to be attributable toischaemia.14 71 Theories have been advanced thatnecrosis and inflammation in choroidal melanomaare indicative of the host's attempt to destroy thetumour. It is of relevance that variable degrees ofapoptosis (shrinkage necrosis) and coagulativenecrosis were observed in 6 of the 21 cases ofuveal melanoma, and, since the lymphocyte trans-formation test was positive in all of these cases, it istempting to conclude that the necrosis was immuno-logically induced. Alternatively, a positive immuneresponse is more likely when a melanoma becomesnecrotic and releases large amounts of tumour-associated antigens. Since it is known that tumourcells which infiltrate and produce metastases aredifferent from the primary tumour mass not only intheir histological characteristics and karyotypicprofiles but also in their susceptibility to cytotoxicdrugs due to surface membrane changes and acquisi-tion of stronger electronegative charge,72 it is tempt-ing to postulate that in those cases of uveal mela-noma in which there was a marked lymphoproli-ferative response the primary tumour contained asmall pre-existing population of cells which waseither more antigenic or was capable of extendingoutside the intraocular environment to mount animmune response.While this may be true, there is no doubt that

infiltration and metastases (in malignant melanoma)are the net results of a variety of cellular charac-teristics including antigenicity, susceptibility toimmune-mediated lysis, anchorage-independentgrowth, multichromosomal abnormality, and therate of growth.72The degree of pigmentation and the extent of

mitotic activity and their correlation with prognosishave been well documented,9 73 74 but in our studywe did not find any significant correlation between

the degree of pigmentation and mitotic activity withthe intensity of the lymphoproliferative response. Itappears that a positive lymphoproliferative responseto melanoma antigens is obtainable in about halfof the patients (i.e., 12 out of 25) with malignantmelanoma of the uvea. However, the lymphoproli-ferative response evoked by melanoma tissue wasmuch weaker than the response seen in culturesstimulated by PHA and PPD. This may mean thatmelanoma antigens are poorly immunogenic. Thedifferences between positive and negative responsesmay be markedly reduced if the antigen is notprepared correctly and the proper culture techniquesare not followed meticulously. The best results areobtained when the lymphocyte culture is challenged24 hours after incubation, since by this time mostof the suppressor T cells are ineffective. A goodlymphoproliferative response was obtained onlywhen lymphocytes, maintained in a gaseous phaseconsisting of 5% carbon dioxide in air, were har-vested 6 days after the addition of the antigen.Since the test was negative in about half of thehistologically confirmed cases of uveal melanoma,a negative immunological result does not excludethe diagnosis of malignancy. Nevertheless, the factthat we did not obtain any false positive results isencouraging, though the control series consisted ofonly 12 cases; further work is required before aconclusion is warranted.

It seems that the patients do not mount a reason-able degree of cell-mediated immunological responseto melanoma antigens unless the tumour hasextended outside the eyeball and become accessibleto immunologically competent cells. It is not sur-prising, therefore, that, although the response ofthe lymphocytes in these patients was tumour-specific, it was largely ineffective in controllingtumour growth.

This peculiar response of uveal malignancy whencompared with malignant tumours elsewhere in thebody can be explained at least in part by postulatingthat both the afferent and efferent limbs of theimmune response in the eye may be subject to delayand modification because of the absence of lym-phatic channels in intraocular tissues.

We thank Professor A. Garner for his support and MissL. Hart and Mrs C. Thornton for their secretarial assistance.

References

1 Newton FH. Malignant melanoma of the choroid. ArchOphthalmol 1965; 73: 198-9.

2 Ashton N. Primary tumours of the iris. Br J Ophthalmol1964; 48: 650-68.

3 Sunba MSN, Rahi AHS, Morgan G. Tumours of theanterior uvea: I. Metastasizing malignant melanoma ofthe iris. Arch Ophthalmol 1980; 98: 82-5.

588



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/


4 Fuchs E. Uber sarkom der aderhaut nebst bemerkungenuber nekrose der uvea. Albrecht von Graefes Arch KlinOphthalmol 1910; 77: 304-92.

5 Jensen OA, Andersen SR. Spontaneous regression of amalignant melanoma of the choroid. Acta Ophthalmol(Kbh) 1974; 52: 173-82.

6 Furth J. Conditioned and autonomous neoplasms. Areview. Cancer Res 1953; 13: 477-92.

7 Dunphy EB, Forrest AW, Leopold IH, Reese AB,Zimmerman LE. The diagnosis and management ofintraocular melanomas. A symposium. Trans Am AcadOphthalmol Otolaryngol 1958; 62: 517-55.

8 Anderson B, O'Neil J. Malignant melanoma of the uvea.Arch Ophthalmol 1957; 58: 337-47.

9 Rahi AHS. Autoimmune reactions in uveal melanoma.Br J Ophthalmol 1971; 55: 793-807.

10 Wong IG, Oskvig RM. Immunofluorescent detection ofantibodies to ocular melanomas. Arch Ophthalmol 1974;92: 98-102.

11 Oettgen HF. Immunotherapy of cancer. N Engl J Med1977; 297: 484-91.

12 Rahi AHS, Garner A, Malaty AHA. Contractile proteinantigens in the cells of malignant melanoma of thechoroid and their diagnostic significance. Br J Ophthal-mol 1978; 62: 394-401.

13 Malaty AHA, Rahi AHS, Garner A. Ostensible anti-melanoma antibodies in patients with non-malignanteye disease. In: Silverstein AM, O'Connor GR, eds.Immunology and Immunopathology of the Eye. NewYork: Masson Press, 1979: 40: 216-9.

14 Rahi AHS. Immunological aspects of malignant mela-noma of the choroid. Trans Ophthalmol Soc UK 1973;93: 79-91.

15 Char DH, Hollinshead A, Cogan DG, Ballantine E,Hogan MJ, Herberman RB. Cutaneous delayed hyper-sensitivity reactions to soluble melanoma antigen inpatients with ocular malignant melanoma. N Engl J Med1974; 291: 274-7.

16 Johnson MW, Maibach HI, Salmon SE. Skin reactivityin patients with cancer. Impaired delayed hypersensiti-vity or faulty inflammatory response? N Engl J Med1971; 284: 1253-7.

17 Manor RS, Livini E, Joshua H, Ben-Sira I. Inhibitionof macrophage migration by choroidal malignantmelanoma-associated antigens in patients with uvealmelanoma. Invest Ophthalmol Visual Sci 1978; 17: 684-7.

18 Pick E. Mechanism of action of migration inhibitorylymphokines. In: Cohen S, Pick E, Oppenhein JJ, eds.Biology of the Lymphokine. London, New York: Aca-demic Press, 1979: 59-1 19.

19 McCoy JL, Dean JH, Cannon GB, Oldham RK, Herber-man RB. Detection of cell-mediated immunity againsttumour associated antigens of human breast carcinomaby migration inhibition and lymphocyte stimulationassays. In: Wyhman J, Staquet MJ, eds. Clinical TumourImmunology. Oxford: Pergamon Press, 1976: 77-86.

20 Field EJ, Caspary EA. Lymphocyte sensitization: an'in vitro' test for cancer? Lancet 1970; fi: 1337-41.

21 Rahi AHS, Otiko G, Winder AF. Evaluation of macro-phage electrophoretic mobility (MEM) test as an indi-cator of cellular immunity in ocular tumours. Br JOphthalmol 1976; 60: 589-93.

22 Goldston AH, Kerr EJL, Lorna, Irvine WJ. The macro-phage electrophoretic migration test in cancer. Clin ExpImmunol 1973; 14: 469-72.

23 Pritchard JAV, Moore JL, Sutherland WH. Joslin CAF.Evaluation and development of macrophage electro-phoretic mobility (MEM) test for malignant disease.Br J Cancer 1972; 27: 1-9.

24 Dale MM, Mongar, J. Role of immunity in diagnosis of

human cancer. Br J Cancer 1973; 28 (supp. 1): 237-42.25 Lewkonia RM, Kerr EJL, Irvine WJ. Clinical evaluation

of the macrophage electrophoretic mobility test forcancer. Br J Cancer 1974; 30: 532-7.

26 Crozier EH, Hollinger ME, Woodend BE, RobertsonJH. An assessment of the macrophage electrophoreticmobility test (MEM) in cancer diagnosis. J Clin Pathol1976; 29: 608-9.

27 Gass JDM. Problems in the differential diagnosis ofchoroidal nevi and malignant melanomas. Am J Ophthal-mol 1977; 83: 299-323.

28 Ferry AP. Lesions mistaken for malignant melanoma ofthe posterior uvea: a clinicopathologic analysis of 100cases with ophthalmoscopically visible lesions. ArchOphthalmol 1964; 72: 463-9.

29 Shields JA, Zimmerman LE. Lesions simulating malig-nant melanoma of the posterior uvea. Arch Ophthalmol1973; 89: 466-71.

30 Harry J. Errors in diagnosis. Symposium on managementof melanoma of the uveal tract. Trans Ophthalmol SocUK 1973; 93: 93-102.

31 Jehn UW, Nathanson L, Schwartz RS, Skinner M.In vitro lymphocyte stimulation by a soluble antigen frommalignant melanoma. N Engl J Med 1970; 283: 329-33.

32 Meltzer MS, Leonard EJ, Rapp HJ, Barson T. Tumour-specific antigen solubilized by hypertonic potassiumchloride. J Natl Cancer Inst 1971; 47: 703-9.

33 Lowry OH, Rosebrough NJ, Farr AL, Randall RJ.Protein measurement with folin-phenol reagent. J BiolChem 1951; 193: 265-75.

34 Carpentier NA, Lange GT, Fiere DM, et al. Clinicalrelevance of circulating immune complexes in humanleukaemia. Association acute leukaemia of the presenceof immune complexes with unfavourable prognosis.J Clin Invest 1977; 60: 874-84.

35 Bloch-Stacher N: Hirschhorn K. Stimulation of non-immunized human lymphocytes by antigen-antibodycomplexes. Proceedings of the third annual leukocyteconference, University ofIowa. Pieke WO, ed. New York:Appleton Century Crofts, 1969: 557.

36 Federman JL, Lewis MG, Clark WH. Communications:Tumour-associated antibodies to ocular and cutaneousmalignant melanomas: Negative interaction with normalchoroidal melanocytes. J Nati Cancer Inst 1974; 52:587-9.

37 Sell S. Tumour immunity. Relevance of animals modelsman. Hum Pathol 1978; 9: 63-9.

38 Federman JL, Lewis MG, Clark WH, Egerer I, SarinLV. Tumour-associated antibodies in serum of ocularmelanoma patients. Trans Am Acad Ophthalmol Otolaryn-gol 1974; 78: 784-94.

39 Muir VY, Turk JL, Hanley HG. Comparison of allergicaspermatogenesis with that induced by vasectomy. I.In vivo studied in the guinea-pig. Clin Exp Immunol1976; 24: 72-80.

40 Colvin RB, Dvorak HF. Role of the clotting system incell mediated hypersensitivity. J Immunol 1975; 114:377-87.

41 Thompson RA. Lymphokine production. In: ThompsonRA, ed. Techniques in Clinical Immunology. Oxford,London: Blackwell Scientific Publication, 1977: 196-200.

42 Houk JC, Sharma VK, Cheng RF. Fibroblast cheloneand serum mitogen (antichalone). Nature (New Biology)1973; 246: 111-3.

43 Tanapatchaiyapong P, Zolla S. Humoral immuno-suppressive substances in mice bearing plasmocytoma.Science 1974; 186: 784-50.

44 Whittaker MG, Rees K, Clark CG. Reduced lymphocytetransformation in breast cancer. Lancet 1971; i: 892-3.

45 Knocke J Kanadie A, Powell AE. Modulation of

589



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/


lymphocytic responses by factors in human plasma.In Arch Allergy 1974; 46: 584-99.

46 Murgita RA, Tomasi TB, Jr. Suppression of the immuneresponse by alpha feto-protein. II. Effect of MLR andmitogen reactions. J Exp Med 1975; 141: 440-52.

47 Cooperband SR, Nimberg R, Schmid K Mannick JA.Humoral immunosuppressive factors. Transplant Proc1976; 8: 225-42.

48 Zembala M, Mytar B, Popiela T, et al. Depressed invitro peripheral blood lymphocyte response to mitogensin cancer patients: The role of suppressor cells. Int JCancer 1977; 19: 605-13.

49 Char DH, Christensen M, Goldberg L, Stein P. Immune-complexes in retinoblastoma. Am J Ophthalmol 1978;86: 395-9.

50 Dernouchamps JP, Vaerman JP, Michiels J, Masson PL.Immune complexes in the aqueous humor and serum.Am J Ophthalmol 1977; 84: 24-31.

51 Oppenheim JJ, Schecter B. Lymphocyte transformation.In: Rose NR, Friedman H, eds. Manual of ClinicalImmunology. American Society of Microbiology, 1976:81-94.

52 Doberson MJ, Bell AM, Jenson AB, Notkins AL.Detection of antibodies to islet cells and insulin withparaffin-embedded pancreas as antigen. Lancet 1979; ii:1078.

53 Schumm DE, Billmire DF, Morris HP. Inhibition ofmitogen-induced lymphocyte blastogenesis by serumfrom tumour-bearing animals: changes with tumourgrowth. Eur J Cancer 1976; 12: 689-94.

54 Nagel GA, Piessens WF, Stilmant MM, Lejeune F.Evidence for tumour-specific immunity in humanmalignant melanoma. Eur J Cancer 1971; 7: 41-47.

55 Golub SH, Morton DL. Sensitization of lymphocytesin vitro against human melanoma-associated antigens.Nature 1974; 251: 161-3.

56 Eilhilali MM, Britton S, Brosman S, Fahey JL. Criticalevaluation of lymphocyte function in urological cancerpatients. Cqncer Res 1976; 36: 132-7.

57 Vanky F, Stjernsnard J, Nilsonneu. Cellular immunityto human sarcoma. J Natl Cancer Inst 1971; 46: 1145-51.

58 Jensen OA. Malignant melanomas of the uvea in Den-mark, 1943-1952; a clinical, histopathological and prog-nostic study. Acta Ophthalmol (Kbh) 1963; Suppl 75:214.

59 Paul EV, Parnell BL, Fraker M. Prognosis of malignantmelanomas of the choroid and ciliary body. Int Ophthal-

mol Clin 1962; 2: 387-402.60 Grierson I, Lee Wr. Erythrocyte phagocytosis in the

human trabecular meshwork. Br J Ophthalmol 1973; 57:400-15.

61 Kerbel RS. Immunologic studies of membrane mutantsof a highly metastatic murine tumour. Am J Pathol1979; 97: 609-22.

62 Nelson DS. In: Nelson DS, ed. Immunobiology of theMacrophage. New York: Academic Press, 1976: 633.

63 Hamlin IME. Possible host resistance in carcinoma of thebreast: a histological study. Br J Cancer 1968; 22: 383-401.

64 Tritsch H. Untersuchungen uber Relationen Zwischenentzundlicher Infiltration, Zell type und prognose beimMalignen Melanoma. Arch Derm Forsch 1972; 244:222-30.

65 Karesen R. The immune reaction against malignantmelanoma studied in a biopsy material. Acta PathMicrobiol Scand 1974; 82: 116-26.

66 Roubin R, Cesarini J-P, Fridman WH, Pavie-Fisher J,Peter HH. Characterization of the mononuclear cellinfiltrate in human malignant melanoma. Int J Cancer1975; 16: 61-73.

67 Fujimoto S, Greene M, Sehon AH. ImmunosuppressorT cells and their factors in tumor-bearing hosts. Suppres-sor Cells in Immunity: International Symposium. SinghalSK, Sinclair NR, eds. London, Ontario: Universityand Western Ontario Press, 1975; 136-48.

68 Broder S, Waldmann TA. The suppressor-cell networkin cancer (Part 1). N Engl J Med 1978; 299: 1281-4.

69 Broder S, Waldman TA. The suppressor cell networkin cancer (part II). N Engl J Med 1978; 1335-41.

70 Weiss L. Dynamic aspects of cancer cell population inmetastasis. Am J Path 1979; 97: 601-8.

71 Reese AB, Archila EA, Jones IS, Cooper WS. Necrosisof malignant melanoma of the choroid. Am J Ophthalmol1970; 69: 91-104.

72 Fidler IJ, Cifone MA. Properties of metastatic and non-metastatic cloned subpopulations of ultraviolet-light-induced murine fibrosarcoma of recent origin. Am JPathol 1979; 79: 633-48.

73 Callender GR, Wilder HC, Ash JE. Five-hundredmelanomas of the choroid and ciliary body followedfive years or longer. Am J Ophthalmol 1942; 25: 962-7.

74 Rahi AHS, Agrawal PK. Prognostic parameters inchoroidal melanomata. Trans Ophthalmol Soc UK 1977;79: 368-72.

590



j.com/

Br J O


ugust 1980. Dow

nloaded from

http://bjo.bmj.com/

Lymphoproliferative response as an uveaLymphoproliferative response as an index ofcellular immunity...

Documents

Transcript of Lymphoproliferative response as an uveaLymphoproliferative response as an index ofcellular immunity...