VIRUS-SPECIFIC T-CELL BANKS FOR “OFF THE SHELF” ADOPTIVE THERAPY OF REFRACTORY INFECTIONS

Abstract

Adoptive immunotherapy with transplant donor-derived virus-specific T-cells has emerged as a potentially curative approach for the treatment of drug refractory EBV+lymphomas as well as CMV and adenovirus infections complicating allogeneic hematopoietic cell transplants. Adoptive transfer of HLA partially matched virus-specific T-cells from healthy third party donors has also shown promise in the treatment of these conditions, with disease response rates of 50–76% and strikingly low incidences of toxicity or GVHD recorded in initial trials. In this review, we examine the reported experience with transplant donor and third party donor-derived virus specific T-cells, identifying characteristics of the viral pathogen, the T-cells administered and the diseased host that contribute to treatment response or failure. We also describe the characteristics of virus-specific T-cell lines in our center’s bank and the frequency with which in vitro culture promotes expansion of immunodominant T-cells specific for epitopes that are presented by a limited array of HLA alleles that are highly prevalent which facilitates their broad applicability for treatment.

INTRODUCTION

In 1991, Riddell et al ^{1, 2} first demonstrated that adoptive transfer of cloned transplant donor-derived T-cells sensitized in vitro with autologous CMV infected fibroblasts could prevent CMV infections in allogeneic HSCT recipients without causing GVHD. Thereafter, adoptive transfer of unselected donor leukocytes containing EBV-specific T-cells ³ or in vitro selected EBV-specific T-cells ⁴ were shown to be effective for treatment or prevention of EBV lymphomas complicating HCT. Since then, several phase I and phase II trials of adoptive immunotherapy with transplant donor-derived EBV, CMV or adenovirus specific T-cells have confirmed these findings and illustrated the potential of this approach in the treatment of viral infections failing standard therapies. For EBV lymphomas and lymphoproliferative diseases, these studies have employed T-cells expanded in vitro following sensitization with autologous B cells transformed with strain B95.8 of EBV. For CMV and adenovirus infections, T-cells can be sensitized by a variety of antigen presenting cells including autologous PBMC, monocyte derived dendritic cells or EBV-transformed B-cells loaded with infected cell lysates or pools of synthetic peptides that span the sequences of immunogenic viral proteins such as CMVpp65 and CMV, Immediate early Antigen-1 (IE-1), or adenovirus hexon^5–13. Alternatively, these antigen presenting cells can be transduced to express one or more immunogenic viral proteins.^{14, 15} The T-cells, usually generated over 4–5 weeks of culture, are enriched for virus-specific T-cells and depleted of alloreactive T-cells.⁶ As a result, adoptive transfer of such cells has resulted in clearance of virus and control of infection in a high proportion of cases, without early toxicities or GVHD. ^5–13 More recent studies have used T-cells isolated directly from donor leukocytes on the basis of their binding viral peptide/HLA tetramers or dissociable streptamers, or on expression of activation markers or cytokines after short-term in vitro sensitization.^16–21 These latter approaches are rapid but yields of virus-specific T-cells are limited and therapeutic doses may not be achieved if specific T-cell frequencies in the donor are low. Nevertheless, following adoptive transfer of doses as low as 10⁴ virus-specific T-cells/Kg, the T-cells expand and can control infection in a high proportion of cases with a low associated incidence of GVHD. Results of studies evaluating adoptive transfer of virus-specific transplant donor-derived T-cells in the treatment of EBV lymphomas have been extensively reviewed previously.^{22, 23} Results of recent trials of transplant donor-derived CMV-specific T-cells for the treatment of CMV infections or viremias persisting despite antiviral drugs are summarized in Table 1.

Table 1.

Clinical Trials of Transplant Donor-Derived CMV Specific T-cells

Authors	Antigens	Mode of Treatment	Responses	Complications	Response Rate
Prophylactic/Preemptive
Riddell, S. et al. 1995 1	CMV Infected Fibroblasts	14 Prophylaxis	No CMV Infections		100%
Peggs, L. 2003 8	CMV Lysate	16 Pre-emptive	8 Cleared, without antivirals 8 Cleared with GCV 2 late Reactivation		50%
Peggs, L. 2011 19	CMV Peptides	7 Prophylactic	No Infections	3 GVHD	100%
	IFN-Capture	11 Pre-emptive	9 Cleared, with GCV		88%
Blyth, E. 2013 11	CMV9965 Vector-Modified DCs or Peptide Pool	29 Prophylactic			83%
		21 Pre-emptive			62%
Leen, A. 2013 57	CMVpp65 Transduced APCs	11 Prophylactic	3 Reactivation, Cleared		73%
Therapy for Persistent Viremia/Disease
Einsele, 2004 7	CMV Lysate	8 Persistent Viremia	7 Cleared off antivirals		87%
Feuchtinger, 2010 9	CMVpp65 Peptide Pool	18 Persistent Refractory Viremia	15 Cleared or Viremia Reduced		83%
Cobbold, 2005 16	CMVpp65 Tetramer	9 Pre-emptive	7 Cleared		78%
	Isolation	2 Persistent Viremia	1 Cleared		50%
Schmitt, 2011 18	CMVpp65 Streptramers	2 Persistent Viremia	2 Cleared		100%
Lucas, K. 2012 12	CMVpp65 Peptide Mix	7 Persistent Viremia	3 Cleared, 1 Reduction		57%
Koehne, G. 201513	CMV Peptide Mix	11 Persistent and Refractory Viremia	11 Cleared		100%
		5 Refractory CMV Disease	3 Cleared		60%

CELLULAR INTERACTIONS CONTRIBUTING TO RESPONSE OR FAILURE OF ADOPTIVE IMMUNOTHERAPY WITH TRANSPLANT DONOR-DERIVED VIRUS-SPECIFIC T-CELLS: LESSONS FROM CLINICAL TRIALS AND MODEL SYSTEMS

Although still limited, the results of clinical trials evaluating adoptive T-cell therapies for EBV, CMV and adenovirus infections post HCT have demonstrated several common landmarks that distinguish clinical and virological responses from treatment failures. Principal among these is the consistent correlation observed between in vivo expansion of the transferred T-cell populations and clinical response. This expansion is demonstrated by 2 to 3 log₁₀ increases in the frequency of the transferred virus-specific T-cells in the patient’s blood and is usually first detected 10–14 days after T-cell infusion.^{6, 7, 24} In contrast, in patients who fail to respond, such increments in virus-specific T-cell frequencies are rarely if ever detected. Strikingly, the doses of T-cells administered have thus far not been found to be a major determinant of response. In part, this may reflect the limited range of doses (10⁴ – 10⁶/Kg) administered. Nevertheless, even at the lowest of these doses, clearance of virus and resolution of clinical disease has been achieved, but only when exponential expansions of the populations of transferred T-cells have also been detected.^{16, 19}

Unfortunately, during this period of early T-cell expansion, clinical and virological parameters may not provide consistent indicators of response. For example, in patients with EBV lymphomas involving the gastrointestinal tract and mesenteric nodes who have failed treatments with Rituximab, circulating levels of EBVDNA may be low or absent despite the presence of growing lesions. Conversely, in patients treated for CMV and adenovirus, levels of viral DNA in the blood often increase transiently shortly after T-cell infusion, ^{8, 12, 19} potentially reflecting lysis of infected cell targets by the infused T-cells. Thus, there is a continuing need for new diagnostic approaches permitting early detection and prediction of patients who are responding to T-cell therapy.

Characteristics of the donor T-cells transferred, the infected cells and the host environment all affect the proliferative potential and effector function of the donor T-cells in vivo.

THE DONOR T-CELLS

In a primate model, adoptive transfer of cloned central memory CD8⁺ T-cells have been found to induce more durable responses than effector memory CD8⁺ T-cells.²⁵ In populations of virus-specific T-cells transferred after immediate isolation of virus-antigen specific T-cells from PBMC, T_CM are relatively enriched, ^{26, 27} which may contribute to the striking level of proliferation observed following their transfer. However, although CMV and EBV-specific T-cells generated in vitro over periods of 4–7 weeks usually contain >95% T_EM, rapid overall expansions in vivo of 4–5 log₁₀ orders of magnitude are regularly observed.⁶ Furthermore, Heslop et al ⁵ have shown that EBV-specific T-cells transduced to express a neomycin resistance gene as a reporter, can be detected up to 10 years post adoptive transfer. To what degree these long-lived virus-specific T-cells are derived from differentiated effector memory T-cells (T_EM) or from Central memory T-cell (T_CM )²⁵ precursors present at the onset of in vitro culture remains to be determined. Recent evidence also suggests that the proliferative potential and persistence of CD8+ T_EM may be sustained by CD4⁺ helper T-cells in the inoculum.^{28, 29} The role of CD4⁺ T-cells as effector cells in the clearance of virus-infected host cells is still unclear. However, HLA-DP restricted adenovirus-specific CD4+ T-cells have been implicated as the principal responders detected in the blood of healthy seropositive donors with controlled latent adenovirus infections.³⁰ Furthermore, patients treated for EBV lymphomas and CMV infection with lines containing >90% CD4⁺ virus-specific T-cells have fared as well as the large population of patients treated with lines containing a predominance of CD8 T-cells (personal observation).

THE VIRUS IN INFECTED OR TRANSFORMED CELLS IN THE PATIENT

The T-cells transferred must be sensitized to peptide epitopes expressed by the virus-infected or transformed cells in the host. This may not be the case if the virus-encoded proteins or processed peptide epitopes used to sensitize the T-cells in vitro induce the generation of T-cells specific for epitopes not expressed by the patient’s endogenous virus. For example, in three patients described by our group and one patient from Baylor University ^{6, 31} who failed to respond to EBV-specific T-cells, the donor T-cells sensitized in vitro with autologous B-cells transformed by the standard laboratory strain of EBV, B95.8, were unable to recognize or kill donor-type B-cells grown from the patient’s donor origin lymphoma that were transformed by the host’s endogenous strain of EBV. On the other hand, T-cells sensitized with donor type B-cells transformed with the patient’s endogenous strain of EBV were able to recognize and kill B-cells transformed with either the endogenous or the B95.8 strain of EBV.⁶ Analysis of the patient in the Baylor series by Gottschalk et al ³¹ demonstrated that the endogenous EBV strain had a mutation in EBNA-3 which made it unrecognizable by the donor derived in vitro generated T-cells. The mutation deleted the sequences of two peptide epitopes of EBNA 3B presented by HLA A1101 targeted by the immunodominant T-cells that had been generated, in vitro, in response to B-cells transformed with B95.8.

EBV and CMV also encode proteins and microRNAs that can impair recognition of infected cells or blunt T-cell responses generated. For example, EBV, in its latent form, downregulates expression of all but 8 of the 80 proteins contributing to generation of lytic virus.³² Furthermore, amino acid sequences within one latent protein, EBNA-1, impair processing and TAP mediated transport of its peptides to class I HLA alleles.³³ EBV latent membrane protein 1 (LMP-1) also stimulates upregulation of Bcl-2 and other proteins that inhibit apoptosis.^{32, 34} In addition, EBV transformed cells express cytokines such as IL-10, IL-13 and TGFβ that inhibit effector T-cell generation and function.³⁴ Similarly, CMV generates a series of evasins, such as US 2, 3, 6 and 11 that can disrupt, in an allele selective manner, the stability and membrane localization of MHC class I molecules, thereby reducing their expression.^{35, 36} One of these evasins, US 3, also disrupts the assembly of HLA class II molecules and the transport and loading of antigenic peptides presented by these alleles.³⁷ Recent evidence also indicates that a microRNA, US 4-1, downregulates an aminopeptidase in the endoplasmic reticulum (GRAP-1) essential to the editing of immunogenic viral peptides prior to their transport and loading on HLA class I molecules for presentation to T-cells.³⁸ Further understanding of these mechanisms will likely lead to approaches that will limit viral evasion and enable improved anti-viral T-cell effector function.

THE IMPACT OF IMMUNODOMINANCE IN HLA DISPARATE HOSTS

HLA disparate HCT donor-derived virus-specific T-cells may also fail to recognize, proliferate and kill virus-infected cells of host origin if they are restricted by HLA alleles not shared by the patient. Such failures result from the fact that virus-specific T-cells in the blood of latently infected donors are usually reactive against a limited number of immunodominant epitopes presented by 1–2 HLA alleles. For example, T-cell cultures from donors expressing HLA B0702, that are generated in response to autologous DCs loaded with a pool of overlapping peptides spanning the sequence of CMVpp65 will invariably, selectively contain T-cells specific for one or two peptide epitopes presented by this allele to the exclusion of all others.^39–41 If the recipient shares this allele, the T-cells will induce clearance of infected cells; however, if the recipient does not, the T-cells are ineffective. We have recently reported this scenario in two patients with host-origin EBV lymphomas who failed to respond to T-cells from HLA disparate donors that were found to be restricted by HLA alleles not shared by the EBV lymphoma.⁶ One of these patients was subsequently treated with EBV-specific T-cells from a third party donor restricted by an HLA allele expressed by the lymphoma and achieved a durable complete remission.⁶

THE INFLUENCE OF THE HOST ENVIRONMENT

Other features of virus-infected cells or the host environment may also limit the proliferation, survival and function of adoptively transferred T-cells. For example, patients who have been previously allosensitized and have circulating anti-HLA antibodies or sufficient residual T-cell function to mount an alloreactive T-cell response could reject the T-cells early after transfer. In patients with massive burdens of virus transformed or infected cells, the expansion of transferred T-cells is also inconsistent, their expansion potentially compromised by antigen-induced apoptosis. Contemporaneous treatment with glucocorticosteroids can also suppress or eliminate the adoptively transferred T-cells. The doses that prevent or restrict growth of these cells in vivo are poorly defined, but doses of 0.5mg prednisone/Kg/day or its equivalent have been suggested as the upper limit tolerable by most centers.⁴² In contrast, we and others have shown that, concurrent treatment with sirolimus or calcineurin inhibitors has not prevented expansion of transferred T-cells or their potential to clear infection, ⁶ potentially reflecting their limited effects on memory T-cells.⁴³

ADDITIONAL LIMITATIONS TO THE TIMELY APPLICATION AND EFFECTIVE USE OF TRANSPLANT DONOR-DERIVED VIRUS SPECIFIC T-CELLS FOR ADOPTIVE IMMUNOTHERAPY

Although adoptive immunotherapy with transplant donor-derived virus-specific T-cells has shown striking potential in the treatment of EBV⁺ lymphomas failing to respond to Rituximab or chemotherapy and drug resistant CMV and adenovirus infections, this approach is currently not available to the majority of affected patients. The principal constraints to the use of transplant donor-derived T-cells are logistic in nature. In vitro generation of EBV, CMV or adenovirus-specific T-cell lines take 40 to 60 days, including the time required to isolate and mature antigen-presenting cells such as cytokine activated monocytes, dendritic cells or EBV-transformed B-cell lines and the 28–35 days required to expand virus-specific T-cells and delete alloreactive T-cell populations capable of inducing GVHD. Because of the acuity and rapid progression of EBV lymphomas ⁴⁴ as well as CMV and adenovirus infections that do not respond to treatment ^{45, 46} such manufacturing times preclude prompt and effective therapeutic intervention unless T-cells are generated prior to onset for patients at risk.

Approaches to overcome such time constraints for T-cell generation include either direct isolation of virus-specific T-cells from leukapheresis products by immunoadsorption of tetramer or streptamer binding T-cells ^16–18 or early selection of T-cells responding to viral antigens based on their expression of activation markers or generation of cytokines. ^19–22 However, if virus-specific T-cell frequencies in the donor’s blood are low, direct selection may not yield significant populations of virus-specific T-cells. Furthermore, while selection with tetramers and streptamers is rapid and efficient for T-cells responding to highly immunogenic viral peptides presented by prevalent HLA alleles such as HLA A0201 or B0702, such epitopes of CMV, EBV and Adv have been identified for only a limited number of HLA alleles. Furthermore, among individuals inheriting less prevalent HLA alleles, such as B3501, the immunodominant T-cell responses in the circulation may not be specific for epitopes presented by this allele, and would not be selectable using HLA B3501-peptide tetramers. Genetic variants of prevalent HLA alleles also differ in their capacity to present specific viral epitopes. For example, T-cells generated from donors expressing HLA A0201, 0205 or 0206 respond to the highly immunogenic CMVpp65 peptide, NLV, and will bind tetramers containing NLV bound to HLA A0201. However, individuals inheriting HLA A0202 or A0211 do not respond to this epitope.⁴⁷

The other major constraint to the application of transplant donor-derived T-cells is the immune status of the donor. If the donor is not immune, virus-specific T-cells cannot be readily generated. This applies to the up to 50% of related and unrelated donors adult donors who are CMV-seronegative and nearly all cord blood transplants. While CMV-specific T-cells can be generated from a proportion of cord blood transplants, ⁴⁸ the epitope specificities of these T-cells differ from those targeted by T-cells from adult donors.⁴⁹ Furthermore, their activity against virus infected targets in vivo may be limited.⁵⁰

The availability of the donor may also be limiting. Some donors may be unwilling or unable to provide the secondary donations of blood or leukocytes needed to generate virus-specific T-cells in time to treat a severe infection.

ADOPTIVE THERAPY WITH HLA-PARTIALLY MATCHED VIRUS-SPECIFIC T-CELLS FROM AN ALLOGENEIC DONOR OTHER THAN THE DONOR OF A PATIENT’S TRANSPLANT

To address several of these limitations and particularly to improve access to virus-specific T-cells for therapy, our own and certain other transplant centers have been exploring adoptive transfer of HLA-partially matched virus-specific T-cells derived from third party donors, that is, healthy individuals other than the donor of the patient’s transplant. The potential advantages and disadvantages of pre-screened and pre generated third party donor-derived versus transplant donor-derived virus-specific T-cells are summarized in Table 2 and will be discussed in light of early trials reported to date. These results, reported from 5 centers, are summarized in Table 3.

TABLE 2.

COMPARISON OF POTENTIAL ADVANTAGES AND DISADVANTAGES OF TRANSPLANT DONOR-DERIVED VERSUS BANKED THIRD PARTY DONOR-DERIVED VIRUS SPECIFIC T-CELLS.

SOURCE OF VIRUS-SPECIFIC T-CELLS
		HCT DONOR	THIRD PARTY DONOR
A	ACCESSIBILITY OF VIRUS-SPECIFIC T-CELLS
	SEROPOSITIVE HCT DONOR	EARLY, LOW YIELD, IF DIRECTLY SELECTED LATE (60 DAYS), HIGH YIELD IF GENERATED BY INVITRO	IMMEDIATE, HIGH YIELD
	CMV SERO NEGATIVE HCT DONOR	NOT AVAILABLE	IMMEDIATE, HIGH YIELD
	CORD BLOOD DONOR	NOT CONSISTENTLY AVAILABLE	IMMEDIATE, HIGH YIELD
B	CAPACITY OF VIRUS-SPECIFIC T-CELLS TO LYSE INFECTED CELLS
	IF STANDARD VIRAL EPITOPES ARE EXPRESSED	YES	YES
	IF VIRAL EPITOPE IS MUTATED	NO	NO, BUT CAN CHOOSE T-CELLS OF DIFFERENT SPECIFICITIES AND RESTRICTION FROM THE BANK
C	VIRUS-SPECIFIC T-CELL PERSISTENCE	LONG (UP TO 10 YEARS)	SHORT (14–90 DAYS)
D	RISK OF HCT REJECTION	NO	THEORETICALLY YES, BUT NOT OBSERVED WITH VIRUS-SPECIFIC T-CELLS GENERATED OVER 28–35 DAYS OF CULTURE
E	RISK OF GRAFT VS HOST DISEASE	YES, IF DIRECTLY SELECTED; NO, OR LOW RISK FOR VIRUS-SPECIFIC T-CELLS GENERATED OVER 28–35 DAYS OF CULTURE	NO, OR LOW RISK FOR VIRUS-SPECIFIC T-CELLS GENERATED OVER 28–35 DAYS OF CULTURE

Table 3.

Use of Third Party Virus Specific T-cells in Allogeneic Hematopoietic Cell Transplant

Transplant Center	Method of Selection	Indication	Failed Prior Therapy	N	Degree of HLA Match	Responses
						CR	PR	SD	NR PD
Edinburgh 51, 52	EBVBLCL/Culture of T-cell Line	EBV+ lymphoma polymorphic (N=2)	2 Ris+ Rituxan	2	≥2 HLA alleles	2	0	0	0
		Monomorphic (N=1)	1 Steroid	1	≥2 HLA alleles	0	0	0	0
Karolinska 17	EBV-HLA Pentamer Sort	EBV+ Lymphoma	None	1	Haplo-identical	1	0	0	0
MSKCC 6, 53, 54	EBVBLCL/Culture of T-cell Line	EBV+ Lymphoma Monomorphic	Rituxan	30	≥2 HLA alleles	16	3	1	10
Baylor 57	Transduced EBVBLCL/Culture of T-cell Line	EBV+ PTLD	Rituxan	8	≥1 HLA alleles	3	3	0	2
		EBV Viremia	Rituxan	1	1 HLA allele	0	0	0	1
CMV
Karolinska 17	CMV-HLA Pentamer Sort	Persistent CMV Viremia	Fosc, GCV	4	Haploidentical	2	1	0	1
Tubingen 19	MNC + CMVpp65 Protein IFNγ+ T-cell Selection	CMV Encephalitis	Fosc, CDV, GCV	1	Partial NS	1	0	0	0
		CMV Pneumonia, Colitis	GCV, Fosc, CDV	1	Partial NS	0	0	0	0
MSKCC 40	DCs + CMVpp65 Peptide Pool Culture of T-cell Line	Persistent CMV Viremia, CMV Pneumonia, Colitis, Encephalitis or Chorioretinitis	Fosc, GCV, CDV	34	≥2 HLA alleles	7	14	6	7
Baylor 57	DCs + EBVBLCL Transduced to express CMVpp65/Culture of T-cell Line	CMV Viremia CMV Pneumonia, Colitis or Retinitis	GCV, Fosc, CDV	25	>1 HLA alleles	11	8	0	6
Adenovirus
Karolinska 17	Adv-HLA Pentamer Sort	Adenovirus Viremia	CDV	1	Haploidentical	0	0	0	1
Baylor 57	Adv Hexon + Transduced DC and EBVBLCL + Culture of T-cell Line	Adenoviral Viremia	CDV	10	≥1 HLA alleles	5	5	0	0
		Adenoviral Colitis, Hepatitis or Pneumonia	CDV	8	≥1 HLA alleles	2	2	0	4
Great Ormond Street 21	Hadv-5 Hexon Peptides/IFNγ+ T-cell Selection	Adeno Viremia Refractory to CDV	CDV	2	Haploidentical	1	0	0	1

UNIVERSITY OF EDINBURGH EXPERIENCE

The use of virus-specific T-cells derived from an individual other than the patient’s own cells or those of the donor of the patient’s transplant was pioneered by Haque et al ⁵¹ who used HLA partially matched EBV-specific T-cells derived from healthy EBV seropositive volunteer blood donors to treat 8 patients who developed either an EBV+ polyclonal lymphoid hyperplasia or monoclonal lymphoma following an allogeneic solid organ (N=7) or hematopoietic cell transplant (N=1). These patients were treated with EBV-specific T-cells derived from a bank of 70 EBV-specific T-cell lines generated from healthy blood donors. The third party T-cells were selected on the basis of matching for 1 or more HLA alleles, their specific cytotoxicity against patient derived EBV+ BLCL and their minimum reactivity against patient PHA blasts. Treatment consisted of intravenous infusions of 10⁶ T-cells/Kg given once every 2 weeks, for up to 6 doses. All three patients with polyclonal disease achieved durable complete remissions (CR); one patient with a lymphoblastic lymphoma achieved a PR while the other 4 patients with monoclonal lymphomas, including the HCT recipient, had progression of disease. In a subsequent multicenter phase II trial, ⁵² they treated 33 transplant recipients with EBV+ lymphomas or lymphoproliferative disease including two who had developed polymorphic lymphomas following hematopoietic stem cell transplants. All patients had failed to respond to reductions of immunosuppressive drugs and 13 had failed treatment with rituximab and/or chemo/radiotherapy. Of the 33 patients, 14 achieved a complete remission and 7 achieved partial remissions 3 of which were sustained for >6 months. Thus, either CR or PR was achieved in 64% of the cases and sustained >6 months in 52%. In this series, both of the patients who developed polymorphic EBV+ lymphomas following HCT achieved a durable complete remission. In this study, matching of donor T-cells for more HLA alleles in the recipient, higher proportions of CD4+ T-cells in the infused EBVCTLs and disease limited to a single site were associated with better response. There was also a trend toward improved outcome among patients who had been seronegative pre-transplant, developed EBVPTLD earlier post transplant, or had lesions that were hyperplastic or Hodgkins-like rather than a polymorphic or monomorphic lymphoma by histology.

MEMORIAL SLOAN KETTERING CANCER CENTER EXPERIENCE

In 2010, our group reported two patients who received HLA partially matched third party donor-derived EBVCTLs as treatment for monoclonal EBV+ DLBCL with extra nodal involvement that developed in B cells of donor origin following allogeneic double umbilical cord blood transplants and were not responsive to Rituximab. In these cases, the EBVCTLs infused were selected based on their restriction by HLA alleles shared by the engrafted cord blood and cord blood derived EBV+ lymphoma. Both patients achieved a durable complete remission.⁵³ We subsequently reported 3 additional patients as part of our overall study of 19 patients who developed pathologically confirmed monoclonal EBV+ DLBCL following an allogeneic HCT who had failed treatment with Rituximab and were then treated with transplant donor or third party derived EBV-specific T-cells.⁶ Of these 3 patients, two also achieved a sustained CR; one failed to respond and died of progressive disease. We have now treated 30 recipients of allogeneic HCT with pathologically confirmed and Rituximab resistant EBV+ DLBCL by adoptive transfer of HLA partially matched third party-donor derived EBV-specific T-cells. Of the 30 patients, 16 have achieved CR and 3 a sustained PR, for a combined (CR+PR) response rate of 63%.⁵⁴ In each case, selection of 3^rd party T-cells was based on matching of at least 2 HLA alleles, specific cytotoxicity against EBVBLCL targets that was also restricted by an HLA allele shared by the patient’s lymphoma and lack of reactivity against PHA blasts derived from the patient and/or fully HLA-mismatched allogeneic targets.

Our group is also exploring the use of third party donor derived CMV-specific T-cells that have been sensitized in vitro with autologous dendritic cells loaded with a pool of overlapping 15-mer peptides spanning the sequence of CMVpp65 and propagated for periods of ≥ 28 days to delete alloreactive T-cells.⁵⁵ Patients eligible for this treatment are allogeneic HSCT recipients with CMV infection and/or persistent CMV viremia that has failed to respond to at least 2 weeks of treatment with conventional antiviral drugs. Selection of 3^rd party donor T-cells in this trial has been based on matching the patient for 2 HLA alleles, demonstration of virus-specific HLA-restricted cytotoxic activity against peptide loaded targets sharing HLA allele(s) with the donor and patient, and absence of cytotoxic activity against unloaded patient and/or fully allogeneic targets. Of 34 patients treated and evaluable thus far, 14 have cleared infection and/or viremia. An additional 8 patients have experienced a PR, defined as at least a 2 log₁₀ reduction in the level of CMVDNA detected in the blood. The overall CR+PR rate is 62%.⁵⁶ In our series no recipient of EBV-specific or CMVpp65-specific 3^rd party T-cells has developed graft suppression or failure or either de novo ≥ grade 2 acute GVHD or a flare of pre-existing acute GVHD. One recipient of EBV-specific T-cells had a grade I skin rash which cleared with topical steroid treatment.

BAYLOR UNIVERSITY EXPERIENCE

The Baylor group has explored the use of third party donor-derived T-cells simultaneously sensitized against three viruses, EBV, CMV and adenovirus, and propagated for periods of 21 days or more in vitro to provide expanded populations of virus-specific T-cells and to deplete alloreactive T-cell populations.⁵⁷ Tri-virus specificity is achieved by sensitizing the T-cells initially with irradiated autologous, PBMC derived monocytes transduced with an adenoviral vector encoding both an adenoviral hexon protein shared by most clinical strains of adenovirus as well as CMVpp65. After 10 days, the T-cells begin weekly stimulation with irradiated autologous EBV+ BLCL also transduced to express the adenoviral hexon and CMVpp65. In this multicenter trial, patients eligible included HSCT recipients with CMV or adenovirus infections or viremia who had failed to respond to at least 7 days of treatment with standard antiviral drugs and patients with EBV viremia or PTLD that had failed to respond within 7 days to Rituximab. Patients ineligible included individuals who had received T-cell specific antibodies or donor lymphocyte infusions within 28 days of proposed treatment, had other uncontrolled infections, or had active grade II-IV acute GVHD or GVHD requiring prednisone doses ≥ 0.5mg/Kg/day. Selection of 3^rd party T-cells was also based on specificity for the virus to be treated, restriction by an HLA allele shared by the patient and T-cell donor, and lack of alloreactivity. Eligible patients received one or more doses of 2×10⁷ virus-specific T-cells/m², with secondary doses given at 2 week intervals. As shown in Table 2, CR+PR rates were 74% and 77% for patients with CMV and adenovirus infections respectively, and 66% for patients with EBVPTLD. In this study, only 2/50 patients treated developed de novo grade I acute GVHD; 6 experienced a flare of pre-existing GVHD but only two patients had grade II or III GVHD.

THE KAROLINSKA INSTITUTE EXPERIENCE

The transplant team at the Karolinska Institute has been evaluating virus-specific T-cells directly isolated from leukophereses or blood donations from HLA partially matched third party donors, using HLA pentamers containing well-recognized immunogenic viral peptide epitopes to sort virus-specific HLA restricted T-cells.¹⁷ They have been able to isolate appropriately HLA-restricted virus-specific T-cells in numbers sufficient to be able to administer doses of 1.8–24.6×10⁴ virus-specific T-cells/Kg to patients with refractory CMV, EBV and adenovirus infections. Of 4 patients treated for persistent CMV viremia that failed to respond to antiviral drugs, 2 cleared the viremia and 1 had a 10 fold reduction in the level of CMVDNA (Table 2). The third party donor-derived CMV-specific T-cells increased in frequency in the 3 responders and could be detected up to 90 days post infusion. Similarly, a patient with an EBV lymphoma treated with an infusion of 1.7×10⁴ HLA A0201 restricted EBV-specific T-cells/Kg from an HLA haplotype disparate mother achieved a durable CR. The donor’s T-cells were detectable for 76 days post infusion. In contrast, a patient with an adenoviral infection refractory to cidofovir failed to respond to two doses of 1.8×10⁴ adenovirus-specific T-cells/Kg. In this patient, the T-cells were detected only once, on day 3 following the initial infusion. In this series, no patient developed GVHD following infusions of the pentamer selected T-cells.

UNIVERSITY OF TUBINGEN EXPERIENCE

Using an alternate strategy, the Tubingen group isolated CMVpp65-specific T-cells from peripheral blood mononuclear cells after a 16 hour sensitization with CMVpp65 protein and subsequent immunoadsorption of virus-reactive interferon-secreting T-cells to paramagnetic beads and isolation of these T-cells with the CliniMACS device. In a series of 18 recipients of allogeneic HSCT treated for persistent CMV viremia or clinically overt CMV infection that failed to respond to GCV or foscarnet, 2 received IFNγ+ CMVpp65-specific T-cells from HLA partially matched third party donors.⁹ One patient who received 2.7×10⁴ T-cells/Kg for CMV encephalitis cleared the infection, while the other who was treated with 0.1×10⁴ T-cells/Kg for CMV pneumonia and colitis failed to respond. By comparison, of 16 patients treated in this series with transplant donor derived CMVpp65-specific IFNγ+ T-cells, 14 achieved a CR or PR.

In summary, although the experience with third party virus-specific T-cells is still limited, the results accumulated to date are quite promising. Although trials directly comparing donor and third party T-cells have not yet been reported, in the larger single arm trials from our center and Baylor University, results of treatment with HLA partially matched third party donor derived T-cells have been comparable to those achieved with transplant donor-derived T-cells in the treatment of EBV lymphomas and only slightly inferior in the treatment of drug refractory CMV infections or persistent CMV viremias. In patients treated for EBV lymphomas, the responses have been durable in 84 to 90% of cases. Moreover, in patients treated for CMV infection or persistent viremia, late recurrences have been rare among those who initially achieved a complete remission.

THE CLINICAL ENIGMA OF RESPONSES TO THIRD PARTY DONOR-DERIVED T-CELLS: THEIR DURABILITY

That the adoptively transferred third party T-cells play a major role in the initial responses observed is consistent with the close temporal correlation observed between expansion of the virus specific third party T-cells in the blood as well as their detection at sites of diseases and both the clinical and virological responses observed. However, the observation that responses to third party donor-derived T-cells have been sustained, and have not usually required multiple repeated courses of T-cells is both surprising and unexplained. Transplant donor-derived virus-specific T-cells can engraft and proliferate in vivo in a tolerant environment. Indeed, as previously noted, Heslop et al ⁵ have detected genetically marked transplant-donor derived EBV-specific T-cells as late as 10 years following adoptive transfer. In contrast, third party T-cells, while able to proliferate in vivo for periods of 1–4 weeks after infusion into immunodeficient HSCT recipients, do not achieve durable engraftment. Indeed, in studies employing PCR amplified techniques to sequentially track short-tandem repeats unique to the third party donor in circulating virus-specific T-cells, such T-cells have been detected in responding patients for periods ranging from only 14 up to 90 days post infusion.^{9, 17, 52, 54, 57} Nevertheless, despite their apparently brief tenure in the host, the responses induced are usually sustained even in patients who are still markedly lymphopenic.

The mechanisms contributing to the sustained responses observed are not known. It is possible that the initial transient expansion of virus-specific T-cells regularly observed in responding patients is sufficient to reduce populations of infected cells to a level permitting reestablishment of the poorly defined equilibrium between host and virus that characterizes asymptomatic latent infections. Small numbers of the third party T-cells may persist long enough in sites of infection to sustain this equilibrium until reconstitution of viral immunity by transplant derived T-cells is achieved. Alternatively, since the transferred third party T-cells, like transplant donor-derived T-cells, do accumulate and lyse infected cells at sites of disease but are themselves allogeneic to the transplant and the host, they may attract transplant-derived effectors, such as NK cells and naïve T-cells, to sites of infection, thereby also fostering cross-sensitization of recruited T-cells to viral antigens and stimulating earlier reconstitution of effective viral immunity by transplant-derived effector cells. Ultimately, immunocompetent T-cells of donor origin are critical to sustained responses, since in those patients genetically incapable of generating T-cell responses that we have treated, clinical and/or virologic responses are sustained only as long as the third party T-cells are detected in the blood. Ongoing research examining interactions between third party T-cells and the infected transplant recipient may elucidate the mechanisms contributing to the demonstrated therapeutic potential of third party donor-derived T-cells and define additional attributes of the adoptively transferred T-cells that enhance their potential for long term benefit.

ESTABLISHMENT OF BANKS OF VIRUS-SPECIFIC T-CELLS FROM THIRD PARTY DONORS FOR BROAD APPLICATION OF ADOPTIVE IMMUNOTHERAPY: CURRENT BANKS, THEIR ADVANTAGES AND LIMITATIONS

To date, only a small number of centers have established cryopreserved banks of varying size containing CMV, adenovirus and/or EBV-specific T-cell lines from normal donors that are GMP grade, characterized as to virus-specificity and HLA type and specifically consented for their use in the treatment of eligible recipients of allogeneic transplants. These centers include, but are not limited to, The University of Edinburgh,⁵² Baylor University,⁵⁷ the Karolinska Institute,¹⁷ The University of Tubingen ⁹ and our own, Memorial Sloan Kettering Cancer Center,⁴⁰ each of which has reported results of ongoing trials. An alternate strategy which has been adopted by the U. of Hannover in Germany ⁵⁸ and the U. of Leiden in the Netherlands (Falkenburg, pers. Communication), is the establishment of a registry of healthy volunteers of known HLA type who have been tested and found to have circulating T-cells specific for one or more of these viruses at sustained frequencies sufficient to permit their rapid and direct isolation from a blood or leukapheresis donation for use in adoptive therapy when needed.

At our center, we have established banks of over 300 EBV-specific T-cell lines and 125 T-cell lines specific for CMVpp65 from healthy seropositive allogeneic HCT donors, specifically for adoptive immunotherapy. Each T-cell line in these banks is GMP grade and separately consented for third party use. All CMVpp65-specific T-cells and EBV-specific T-cells are HLA typed at high resolution and characterized as to virus specificity and HLA restriction. All CMVpp65-specific T-cell lines and a growing proportion of the EBV-specific T-cell lines are also characterized as to their peptide epitope specificities.⁴⁰ All lines are negative for alloreactivity or non-specific cytotoxic activity. They are also microbiologically sterile and contain <5 IU of endotoxin/ml dose of T-cells. The T-cells are cryopreserved at doses that permit rapid “off the shelf” use for adoptive immunotherapy. As a result, for patients transplanted at our center, an appropriately HLA-restricted EBV or CMVpp65-specific T-cell line can be rapidly identified and the patient treated within 1 day of referral.

In addition to their immediate accessibility, these banks of third party donor-derived virus-specific T-cells also provide unique advantages for recipients of HLA non-identical HCTs. One advantage is that the banked T-cells are already characterized as to their HLA restriction, which allows selection of T-cells restricted by an HLA allele expressed by the virus-infected or transformed cells in the patient. Indeed, in a survey of consecutive transplant recipients at our center, we found that we could identify CMVpp65-specific T-cell lines that were 2 HLA-allele matched and appropriately restricted for 93% of 137 recipients of HLA non-identical related or unrelated grafts and 98% of 68 recipients of HLA non-identical cord blood transplants.⁴⁰ In contrast, when we examined the HLA-restrictions of CMVpp65-specific T-cells generated from the donors of HLA non-identical related or unrelated HSCT, they were restricted by an HLA shared by the recipient of their transplant in only 60–70% of cases.

A third advantage, which we recently documented, is that certain patients may fail to respond to virus-specific T-cells specific for epitopes presented by one HLA allele shared by the infected cells of the host if the viral epitope is not expressed by the endogenous virus.^{6, 31} Such patients may, however, respond to secondary treatment with T-cells specific for a different epitope presented by another shared HLA allele. For example, in one recipient of a cord blood transplant who developed an EBV+ lymphoma in the cord blood cells, we found that the third party EBVCTL initially infused, while appropriately HLA restricted and cytotoxic against EBVBLCL transformed by the B95.8 strain of EBV failed to respond to the cord blood donor-derived B cell lymphoma cell bearing the virus in the patient. In this case, infusion of a second line specific for an epitope presented by a different HLA allele shared by the lymphoma and cytotoxic against cells transformed by the endogenous strain of EBV induced a durable remission of disease.⁵⁴

FEASIBILITY OF ESTABLISHING BANKS OF VIRUS-SPECIFIC T-CELLS

A major question to be addressed in estimating the feasibility of developing banks of pathogen-specific T-cells is: how large must the bank be for it to be able to provide T-cells for the population of transplant patients with infection? Based on requests for use of these cells, our bank of over 300 EBV-specific T-cells has been able to provide appropriately HLA restricted EBV-specific T-cells that are matched with the patient for at least two HLA alleles for over 98% of patients referred to our center for treatment. The bank of 125 CMVpp65-specific T-cells is also diverse and has permitted selection of appropriately HLA restricted CMVpp65-specific T-cells matched for at least 2 HLA alleles for 93–98% of the patients referred thus far.

In the multicenter trial reported by Haque et al,⁵² a bank of 107 EBV-specific T-cells was also able to provide T-cells matched for 1 class I and an HLA DR allele for 87% of organ allograft recipients with PTLD referred for treatment. Strikingly, using a bank of only 32 tri-virus-specific T-cell lines, Leen et al ⁵⁷ were able to identify suitable HLA-restricted T-cells matched for one or more HLA alleles for 90% of the patients referred for treatment in their multicenter trial. Analysis of our bank of CMV and EBV specific T-cells suggests that a major factor contributing to the broad applicability of T-cell banks of such limited size derives from two distinct but interrelated characteristics of the transplant donors that have donated cells to these banks. First, while this donor population expresses the full panoply of HLA alleles that would be expected in the ethnographic diverse population of New York, those HLA alleles prevalent in all ethnographic groups are also represented at expected higher frequencies. As a result, most patients and their marrow, PBSC or cord blood donors inherit at least one of these prevalent HLA alleles. The second and somewhat unexpected finding, however, is that those viral epitopes eliciting the immunodominant T-cell responses that emerge in the T-cell lines in our bank are also predominantly presented by a limited set of these same prevalent HLA alleles. This was illustrated by an analysis of our bank of 119 CMVpp65-specific T-cell lines.⁴⁰ These lines are generated by sensitizing the donor’s T-cells with autologous dendritic cells loaded with a pool of overlapping 15-mer peptides spanning the sequence of CMVpp65.⁵⁵ This approach permits sensitization of both CD8 and CD4+ T-cells with peptides of appropriate length, thereby potentially eliciting responses to epitopes presented by either class I or class II alleles. Strikingly, despite the fact that 165 different HLA A, B, C and DR alleles are represented in this bank, only 47(28%) present epitopes eliciting CMVpp65-specific T-cells detected in these lines. Furthermore, the immunodominant epitopes in these lines were presented by only 19 HLA alleles. Indeed, T-cells responding to epitopes presented by only 3 HLA alleles, HLA B0702, A0201 and B3501, account for 54% of the T-cell lines in the bank. Furthermore, evidence also indicates that immunodominant T-cells are restricted by a hierarchy of these alleles. Thus, as previously observed by Lacey et al,³⁹ epitopes of CMVpp65 presented by HLA B0702 are consistently dominant. Of 25 donors in our bank inheriting HLA B0702, all exhibited immunodominant T-cell responses to CMVpp65 restricted by this allele.⁴⁰ Similarly, in our bank, the NLV epitope of CMVpp65 presented by HLA A0201 elicited dominant T-cell responses in 30/39 donors inheriting this allele, the exceptions being donors co-inheriting HLA B0702.⁴⁰ Suksdolak et al ¹⁷ in a study of T-cell responses to CMVpp65 in 120 CMV seropositive donors also found that HLA A0201 restricted T-cell responses were dominant in all donors tested except those co-inheriting HLA B07, B13 or A03. More recently, our group has reported evidence that T-cells responding to epitopes presented by this hierarchy may also be more consistently effective following adoptive transfer.⁵⁹ Such findings, drawn from analyses of virus-specific T-cells from normal seropositive donors, provide evidence for a hierarchy of HLA alleles that present epitopes recognized by the immunodominant T-cells that are maintained in the blood of healthy CMV seropositive latently infected individuals. Thus, as further data emerge regarding the hierarchy of T-cell responses to EBV, CMV and other latent viruses and the contribution of immunodominant T-cell responses to sustained, effective immunity, selection of viral epitope-specific T-cell lines for T-cell banks may be further refined so as to insure not only broad applicability but also consistent therapeutic effectiveness.

WHEN TO CONSIDER TREATMENT WITH VIRUS-SPECIFIC T-CELLS FROM A THIRD PARTY DONOR

Because third party donor-derived virus specific T-cells are expected to proliferate for only 1–4 weeks in immuno compromised allogeneic HCT recipients and usually persist for no more than 90 days post infusion, they have thus far been used exclusively for treatment rather than prophylaxis. Furthermore, since first line treatments are available, specifically, Rituximab for EBV lymphoma, ganciclovir, foscarnet or cidofovir for CMV and cidofovir for adenovirus, current Phase II trials evaluating adoptive therapy with third party donor-derived virus-specific T-cells have been focused on patients who are either not responding to first line treatment or cannot tolerate their continued use due to toxicity. This treatment approach is outlined in Figure 1. However, based on the low risk of toxicities or graft versus host disease observed in Phase II trials thus far, our own and other centers are now initiating trials in which third party donor-derived virus-specific T-cells are evaluated as an adjuvant to standard agents in the initial pre-emptive treatment of these infections to determine whether and to what degree such combined treatments will increase the proportion of patients that achieve rapid and sustained clearance of viremia, prevention or consistent resolution of overt clinical disease and prevention of recurrences of infection late after treatment.

FIGURE 1.

TIMING FOR INTRODUCTION OF THIRD PARTY, VIRUS-SPECIFIC T-CELLS IN ADOPTIVE THERAPY FOR HCT PATIETNS LACKING ACCESS TO TRANSPLANT DONOR-DERIVED VIRUS-SPECIFIC T-CELLS

FUTURE CONSIDERATIONS

While current trials of third party donor-derived virus-specific T-cells have been limited to single centers or consortiums of a small number of centers⁵⁷, these trials and the worldwide experience with cord blood transplants⁶⁰ have demonstrated the transportability of cryopreserved cells, and indicate the potential for broad application of banked virus-specific T-cells both nationally and internationally. For example, for the near future, cooperative groups such as the BMTCTN have prioritized multicenter trials of CMV-specific T-cells. Similar studies are being planned in Europe.

The promising results of trials employing banked third party donor-derived EBV and CMV specific T-cells as well as emerging experience with adenovirus-specific T-cells have also stimulated interest in their uses in the treatment of other viruses affecting immunocompromised hosts that have responded to transplant donor-derived T-cells, including HHV-6, the papovaviruses, BK and JC as well as the human papilloma virus.^61–64

Several strategies are also being explored to enhance both the feasibility of generating virus-specific T-cells for the populations in need and insuring their activity in patients treated. To increase feasibility and applicability, Leen et al ^{14, 15, 65} have pioneered the development of T-cells specific for multiple viruses from single donor samples, using T-cells sensitized with antigen presenting cells either transduced with adenoviral vectors or transfected with DNA plasmids to express the most immunogenic proteins of each virus (e.g. EBV transduced B cells expressing CMVpp65 and adenovirus hexon) or loaded with mixed pools of peptides derived from each virus.⁶⁶ While each of these approaches has already demonstrated significant potential, the HLA restriction of T-cells generated against each virus may differ and therefore must be ascertained to insure that the virus-specific T-cells from the HLA partially matched donor are restricted by an HLA allele shared by infected cells in the patient.

An alternate approach being explored at our center is to sensitize each donor’s T-cells with viral peptide loaded artificial antigen presenting cells each expressing a single class I or class II HLA allele together with critical costimulatory proteins.^{67, 68} This approach permits generation of T-cells against multiple viruses, each restricted by the same HLA allele. This approach also permits consistent generation of different lots of T-cells, each restricted by a distinct HLA allele expressed by the donor’s T-cells and specific for either a dominant or subdominant epitope of the virus.

The potential applications of virus-specific, third party donor derived T-cells genetically engineered to express a second, tumor-specific T-cell receptor or chimeric antigen receptors (CAR) are also being considered. Autologous EBV-specific T-modified to express a CAR specific for GD2, have shown promise in the treatment of cell patients with neuroblastoma.^{69, 70} Banked T-cells specific EBV and CMV are depleted of alloreactive T-cell populations that can cause GVHD, and are potent, cytotoxic effectors that can potentially be activated on a sustained basis by engagement of their virus-specific T-cell receptors with latently infected cells resident in the patient. Because of these attributes, adoptive transfer of such cells, modified to express a second, tumor specific TCR or CAR, could have particular advantages in allogeneic marrow main transplant recipients in terms of both safety and potential for more effective and sustained activity against residual leukemia call in the host.

CONCLUSIONS

Adoptive immunotherapy using banked virus-specific T-cells derived from HLA partially matched healthy consenting donors has shown considerable early promise in the treatment of EBV+ lymphomas and infections due to CMV and adenovirus complicating allogeneic hematopoietic cell transplants. Such T-cells are particularly applicable to the treatment of patients receiving transplants from seronegative donors or cord blood transplants. Such banks permit selection of T-cells on the basis of HLA allele phenotype, viral specificity and HLA restriction, which may provide distinct advantages, particularly in the treatment of recipients of HLA non-identical HCT. Early expansion of these T-cells in vivo and their accumulation in responding sites of disease underscore their contribution to the initial clinical and virologic responses observed, but the cellular interactions contributing to the high rate of sustained responses are still unclear. Nevertheless, early results of trials with these “off the shelf” broadly applicable T-cells indicate their therapeutic potential and warrant evaluation in prospective multicenter trials.