Drug Discovery and Therapeutic Delivery for the Treatment of B and T cell Tumors

Abstract

Hematological malignancies manifest as lymphoma, leukemia, and myeloma, and remain a burden on society. From initial therapy to endless relapse-related treatment, societal burden is felt not only in the context of healthcare cost, but also in the compromised quality of life of patients. Long-term therapeutic strategies have become the standard in keeping hematological malignancies at bay as these cancers develop resistance to each round of therapy with time. As a result, there is a continual need for the development of new drugs to combat resistant disease in order to prolong patient life, if not to produce a cure. This review aims to summarize advances in targeting lymphoma, leukemia, and myeloma in both cutting-edge and well established platforms. Current standard of treatment will be reviewed for these malignancies and emphasis will be made on new therapy development in the areas of antibody engineering, epigenetic small molecule inhibiting drugs, vaccine development, and chimeric antigen receptor cell engineering. In addition, platforms for the delivery of these and other drugs will be reviewed including antibody-drug conjugates, micro- and nanoparticles, and multimodal hydrogels. Lastly, we propose that tissue engineered constructs for hematological malignancies are the missing link in targeted drug discovery alongside mouse and patient-derived xenograft models.

Graphical abstract

1. Introduction to B and T cell tumors and need for drug discovery and delivery technologies

Hematological malignancies are lymphoproliferative disorders of B and T cells that comprise of lymphomas, leukemias, and myelomas. Occurrences of these cancers can be attributed to a variety of factors including but not limited to aging, exposure to toxins, radiations, and chemicals, repeated infections and a weakened immune system, as well as side effects of several biologics (e.g. Humira™). Chronic use of immunosuppressive medications for diseases such as inflammatory bowel disease, weakens the host immune response allowing development of lymphoid malignancies¹. Most of the hematological malignancies represent a heterogeneous disease associated with a variety of clinical presentations and genetic diversity². The heterogeneity of these hematological malignancies present unique challenges to conventional cancer treatment. Owing to the inherent heterogeneity of B and T cell tumors³, there are very different responses to treatments. Many patients with these tumors do not show a durable response to treatments, necessitating salvage therapies such as autologous stem cell transplant that often have poor patient outcomes. A myriad of independently predictive biomarkers for resistance have been identified, including gene expression signature, stromal signatures, epigenetic silencing of specific genes, and specific mutations^{4, 5}. But none are sufficient to predict resistance in a given patient and few are helpful in guiding selection of targeted therapies. Therefore, new treatments are needed to improve clinical outcome of tumors of immune cells. Moreover, many of the first-line immunochemotherapy regimens are too toxic to be tolerated by the elderly or by people in the developing world, where infectious diseases are more difficult to manage following treatment. Clearly, novel approaches to tumor treatment are needed across drug discoveries and delivery of therapeutics.

Lymphoma represents an umbrella term applied to over 30 distinct clinical malignancies of B and T cells. While T cell lymphomas are highly prevalent, most lymphomas originate from B lymphocytes, and their diversity can often be traced to the normal precursor B cell⁶. The bulk of these B cells are antigen-experienced germinal center or post-germinal center B cells which undergo massive proliferation, somatic hypermutation, and antibody class switching⁶. Lymphoma originates in the lymph nodes and the lymphatics and comprises of two categories of diagnosis: Hodgkins and Non-Hodgkins. Hodgkins lymphoma is identified by the presence of Reed Sternberg cells, a form of malignant mature B lymphocytes⁷. Non-Hodgkin lymphoma encompasses all other forms of lymphocytic and myeloid cancers with origins in the lymphatics, and includes several subtypes including hairy cell lymphoma, mantle cell lymphoma, diffuse large B cell lymphoma (DLBCL), Burkitt lymphoma⁸, and peripheral T cell lymphomas. DLBCL is the most common lymphoma representing ~30% of all B cell Non-Hodgkin lymphomas, and gene expression profiling has allowed sub-classification into germinal center B cell-like (GCB) DLBCL and activated B cell-like (ABC) DLBCL subtypes. Improved therapies are needed for all DLBCLs but most urgently for ABC-DLBCLs, which are the most chemo-resistant (5-year overall survival ~ 45%^{2, 4}). Peripheral T cell lymphomas are a molecular and clinical heterogeneous group of non-Hodgkin lymphomas that include, as the most prevalent sub-types, the nodal type peripheral T cell lymphoma not otherwise specified, anaplastic large cell lymphoma, and angioimmunoblastic T cell lymphoma. Although some subtypes may follow a more benign prolonged course, the vast majority of T cell lymphoma patients have poor prognoses due to the combination of an aggressive clinical course and the lack of specific treatments^{9, 10}. Most T cell lymphomas have been commonly treated with therapeutic regimens originally designed for B cell lymphomas, resulting in poor efficacy and frequent relapses^{10, 11}. 2016 incident estimates for the United States predicts 136,960 new lymphoid neoplasms with the most significant forms of lymphoma, with regards to prevalence, remain to be small lymphocytic lymphoma (chronic lymphocytic leukemia), mantle cell lymphoma, diffuse large B-cell lymphoma, follicular lymphoma, and multiple myeloma¹².

On the other hand, leukemia are cancers of B cells and other leukocytes with origins in the bone. There are four subgroups of leukemia, defined by the age at onset (acute or chronic) and the category of cell type that is afflicted (myeloid or lymphoid): acute myeloid leukemia (AML), acute lymphoid leukemia (ALL), chronic myeloid leukemia (CML), and chronic lymphocytic leukemia (CLL). Estimates for 2016 predict over 60,000 new cases and over 24,000 leukemia related deaths in the United States alone¹³. While the death rates for acute lymphocytic and chronic myeloid leukemia have significantly dropped since the introduction of Bruton’s tyrosine kinase inhibitors in leukemia treatment¹³, there remains room for drug discovery and development for tyrosine kinase inhibitor resistant leukemias. Finally, a third type of hematological malignancy is Multiple myeloma, which is a cancer of plasma cells. Plasma cells, in the healthy state, mass produce antibodies in the body. Estimates from 2015 predict that the number of myeloma cases in the United States in 2020 will increase to 11,581 cases from the 9,083 cases in 2010¹⁴. In addition, the median survival for multiple myeloma is only 6 years post-diagnosis and there has yet to be a cure for the disease¹⁵. Treatment options vary by patient age, as the median age at diagnosis is 69 years old¹⁵.

With incredible diversity in cancerous cell types and behaviors within lymphoma, leukemia, and myeloma, diverse therapeutic strategies are required for disease control and achieving remission. Current standards of care for common lymphomas, leukemias, and myeloma are summarized in Table 1 by disease progression stage. It should be noted though, that disease treatment strategy also largely depends on other factors like age and comorbidities rather and the stage of cancer progression and these elements are not represented in this table. From therapeutic delivery standpoint, major challenges within the cancer therapeutics lie with innate drug toxicity and barriers against intracellular drug delivery. The scope of this review is to highlight the advancements in lymphoma, leukemia, and myeloma therapies from drug discovery and drug delivery standpoint. We will review biomolecule engineering including antibody based therapies, targeted enzymes, and small molecule inhibitors. In addition, new cell, tissue, and material engineering techniques will be covered with respect to their impact on potential treatments for the hematological malignances highlighted here.

Table 1.

Gold standards of treatment for most common lymphoma, leukemia and myeloma subtypes.

Malignancy	Early stage therapy	Late stage therapy	Therapy after primary refractory	Post remission therapy	Therapy after relapse
Hodgkin lymphoma	ABVD combination chemo with IFRT to lymph nodes223’224	ABVD or BEACOPP225	HDC with ASCT224	–	Treatment with initial chemo- 80% CR rate226; vinorelbine227 or gemcitabine228
DLBCL	R-CHOP229	–	HDC with ASCT in <60 years old230	–	Rituximab, Gemcitabine, and oxiplatin231 HDC with ASCT230
Follicular lymphoma	Radiation therapy232	(when symptomatic) R-CHOP or BR233	R-CHOP234	–	–
AML	Induction therapy of cytarabine and and doxorubicin235; followed by consolidation chemotherapy or allogeneic stem cell transplant236	–	–	Cytarabine consolidation therapy235; Allogeneic HSCT237	Allogeneic stem cell transplant237, 238
ALL	Hyper-CVAD239; Hyper-CVAD and Imatinib in Philadelphia chromosome translocation240	–	–	Allogeneic stem cell transplant239	Hyper CVAD241 high dose cytarabine and anthracycline242; methotrexate and asparaginase243; VAD244; methotrexate and asparaginase245
CML	TKIs: imatinib mesylate246, dasatinib247, or nilotinib248	After resistance to initial TKIs: Bosutinib249, ponatinib,250	Switch to unused TKI251	–	After failure of 2 or more TKIs: allogeneic stem cell transplant252
CLL	FCR253 or BR254 in fit patients; Rituximab255 or obinutuzumab256 plus chlorambucil in unfit patients	–	–	–	Kinase inhibitors: Idelalisib257 or Ibrutinib258; Allogeneic stem cell transplant259; or Alemtuzumab260, 261
Multiple Myeloma	Thalidomide and dexamethasone262, lenalidomide263, bortezomib264	–	ASCT265	–	Bortezomib, thalidomide, or lenalidomide266; ASCT267

Acronyms used: doxorubicin, bleomycin, vinblastine, dacarbazine (ABVD combination chemo); Involved-field radiation therapy (IFRT); bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone (BEACOPP); High-dose chemotherapy (HDC); autologous stem cell transplant (ASCT); complete response (CR); rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone (R-CHOP); Bendamustine plus rituximab (BR); hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone (hyper-CVAD); vincristine, doxorubicin, and dexamethasone (VAD); tyrosine kinase inhibitors (TKIs), fludarabine, cyclophosphamide, rituximab (FCR)

2. Engineered Antibodies, Enzymes, and Antibody-Drug Conjugates to Improve Therapies against B and T Cell Tumors

Engineered therapeutic antibodies as drugs have resulted in substantial benefits to public health, in particular to the area of cancer. To date, more than 20 recombinant therapeutic antibodies have received approval for the treatment of various diseases. These antibodies are mostly of the immunoglobulin G (IgG) class 1. In order to achieve clinical efficacy, two hallmark functionalities have been identified for engineered therapeutic antibodies: target-specific binding by the antigen binding Fab fragment and Fc domain mediated effector functions. The Fc function involves antibody-dependent cell-mediated cytotoxicity and complement-dependent cytotoxicity^{16, 17}. Generally, engineered antibodies eliminate target cancer cells through four separate pathways: antibody-dependent cellular cytotoxicity, antibody-dependent cellular phagocytosis, complement-dependent cytotoxicity, and through the cancer specific delivery of conjugated drug¹⁸. Fc receptors (FcRs) are a group of glycoproteins organized into classes, such as FcαR, FcγR and FcεR; FcγRs are expressed mostly on leukocytes, and are the class of receptors most relevant for the function of therapeutic antibodies. FcRs are the link between antigen-bound antibody and immune effector cell elimination. FcγRs have been categorized according to their affinity for specific IgG subclasses and the type of signaling pathway that they trigger; that is, whether it is inhibitory or activating¹⁹. Fc glycosylation is necessary for therapeutic antibodies to elicit effector functions²⁰. Complement activation by IgG is made possible by asparagine 297 with its associated glycans in the CH2 domain of the antibody²¹. Such glycans are protected within the secondary and tertiary structure, supporting the fact that glycosylation of this area is critical for proper binding to FcγR despite minimal contact between the glycan and its respective receptor²². Along similar lines, manipulation of another saccharide, fucose, has shown to enhance binding of IgG1 to FcγRIIIa leading to greater efficiency in binding and antibody-dependent cellular phagocytosis downstream^23–26

Antibody engineering has been implemented for the treatment of cutaneous T cell lymphoma^{27, 28}, follicular lymphoma^{29, 30}, diffuse large B cell lymphoma³⁰, Hodgkins lymphoma^{31, 32}, and Burkitt’s lymphoma³³. Antibody engineering for hematological malignancies gained popularity after the landmark success of rituximab (sold under the brand name Rituxan), a CD20 targeting, chimeric anti-human monoclonal IgG1 antibody that effectively targets several cell types within the B cell lineage^34,35. Rituximab has achieved clinical success in B cell malignancies including follicular lymphoma^{29, 36} diffuse large B cell lymphoma^{35, 37}, and relapsed CLL³⁸, and has served as a model for the clinical potential of antibody engineering in hematological malignancies. Antibody engineering for lymphoma therapeutics targets CD20^{29, 30, 35, 37, 39–41}, killer cell immunogloblulin-like receptors⁴¹, CD22^{30, 33}, C-C motif chemokine receptor 4 (CCR4)^{27, 28, 42, 43}, CD80²⁹, CD27⁴⁴, CD40³², CD19⁴⁵, CD30³¹, CD70⁴⁶ and even soluble interleukin (IL)-6⁴⁷, IL-2³¹, and IL-12³¹. In the case of B cell malignancies, additional effector mechanisms are present in which direct cytotoxicity occurs from antibody internalization and the triggering of apoptotic or non-apoptotic events⁴⁸. Likewise, Mogamulizumab (KW-0761), an defucosylated humanized monoclonal antibody was found to successfully target C-C motif chemokine receptor 4 (CCR4) in both peripheral and cutaneous T cell lymphomas in CCR4 positive patients in a phase II Japanese study⁴². It is important to consider genetic diversity and polymorphism in designing therapeutic antibodies. Genetic analyses have revealed that lymphoma patients carrying the FcγRIIIa-v158 allotype responded better to treatment with the chimeric CD20-specific IgG1 antibody rituximab than those patients carrying the FcγRIIIa-F158 allotype^{49, 50}. In 2014, Obinutuzumab, also known as GA101 or Gazyva⁵¹, a glycoengineered type II humanized anti-CD20 monoclonal antibody was named a successor to its own long-time blockbuster drug Rituxan (rituximab). Gazyva, a fully humanized antibody, has a dual mechanism of action where the antibody binds and directly kills B cells by introducing caspase-independent programmed cell death, and through antibody dependent cytotoxicity it recruits the immune system to attack B cells. Gazyva is a type II antibody, which binds to a different epitope on CD20, rather than a type I antibody like Rituxan, which engages complement dependent cytotoxicity and antibody dependent cytotoxicity but does not potently induce programmed cell death upon binding to CD20. Finally, the discovery of high expression levels of programmed cell death 1 Ligand 1(PD-L1) in classical Hodgkin’s lymphoma by programmed cell death protein 1 (PD-1), a key immune-inhibitory molecule expressed on T cells has led to the revolutionary field of immunotherapy. It was discovered that PD-L1 expressing cells can escape T-cell-mediated cellular cytotoxicity by exploiting the inhibitory PD-1 immune checkpoint. Therefore, blocking the PD-L1 with therapeutic antibodies that block the PD-1–PD-L1 axis induce durable clinical responses against a growing list of solid tumors as well as against both Hodgkin’s’ and Non-Hodgkin’s lymphomas, including Diffuse large B Cell Lymphomas^52–54.

A more advanced and rapidly emerging use of antibody engineering is in targeted drug delivery of drugs and other molecules. Over the years, antibody engineering has evolved from monoclonal antibody development to the engineering of antibody drug conjugates (ADCs) for targeted, site-specific cancer drug delivery. ADCs are engineered by chemically conjugating single or multiple cytotoxic small-molecule drugs to a surface antigen receptor-targeted antibody^{55, 56}, and function as targeted drug delivery system to guide the toxic conjugated drugs specifically to malignant cells while minimizing potential systemic toxicity. Multiple mechanisms have been exploited for cytosolic delivery. For example, the reducing nature of the cytosol has been used extensively in protein-conjugate chemistry to trigger release of the payload^{56, 57}. Anti-cancer drugs like dimeric pyrrolobenzodiazepine have been conjugated onto monoclonal antibodies to target cancerous cell types⁴⁶. Pyrrolobenzodiazepine combats uncontrolled cell growth by crosslinking double stranded DNA, effectively damaging beyond the ability of cellular DNA repair mechanisms therefore triggering target cell death⁴⁶. Similarly, calicheamicin has also been incorporated into ADCs as the therapeutic component of inotuzumab ozogamicin for follicular lymphoma, DLBCL, and refractory Non-Hodgkin’s lymphomas³⁰. This anti-tumor antibiotic cleaves double stranded DNA with remarkable specificity⁵⁸, and ADCs with this drug have proven to be effective in refractory Non-Hodgkin’s lymphomas in combined treatment with rituximab. Other ADC drugs implemented for the treatment of lymphoma include DM4, a maytan-sinoid that blocks cell division at the microtubule level⁴⁵. For T cell lymphomas, such as the relapsed anaplastic large cell lymphomas, SGN-35 (Brentuximab vedotin, Adcetris^®) has emerged as a new anti-CD30 ADC combination with the potent synthetic drug monomethyl auristatin E. The antibody is conjugated through an enzyme cleavable linker, which makes it stable in the bloodstream but enzymatically cleavable upon internalization into CD30-expressing T cell tumors, releasing monomethyl auristatin E and leading to cell death. For better understanding of new targets of therapy in T-cell lymphomas, readers are recommended to consult the review by Erter et al.⁵⁹

In addition to Non-Hodgkin’s lymphomas, CD20 (rituximab), CD33(gemtuzumab ozogamicin), CD22 (inotzumab ozogamacin and epratuzumab), CD19 (blinatumomab), and CD52(Alemtuzumab/Campath 1-H) have served as targets of antibody therapy for the treatment of acute lymphoblastic leukemia⁶⁰. Recent research efforts have focused on new cellular surface targets like CD27⁴⁴ and EphA3⁶¹, as well as improving cancer eradication though development of antibody drug conjugates for CD22⁶². Anti-CD22 antibody therapies have arguably been the most successful, and clinically translatable in treating ALL and other leukemias and lymphomas. Notable innovations to antibody-drug conjugates have evolved from the success of anti-CD22 monoclonal antibody therapies. Since upon binding with CD22, the antibody therapy is endocytosed, CD22 antibodies are an intriguing drug delivery platform for drugs with intracellular actions. Satake et al. used monoclonal anti-CD22 antibodies to traffic antisense oligonucleotides for inhibiting translation of MYC-associated factor X (MAX) dimerization protein 3 (MXD3), an important transcription factor involved in preB-cell acute lymphoblastic leukemia⁶². Results of their preclinical study showed that the conjugate successfully trafficked the oligonucleotides to the leukemia cells without off target effects and at one-twentieth of the dose required for similar effects from anti-CD22 monoclonal antibody therapy alone^62,63.

Anti-CD33 ADCs have also been developed for the targeting of leukemia cells and multiple generations of CD33 targeted ADCs have evolved, as summarized in Fig. 1 and reviewed in detail elsewhere⁵⁶. Gemtuzumab ozogamicin (GO) is a calichaemicin linked to a CD33-specific human antibody and functions by cleavage of the reductively labile disulfide-based linker⁵⁷. GO was the first antibody-drug conjugate to be approved by the FDA for acute promyelocytic leukemia⁶³. The treatment of other AML subtypes has proved to be less efficacious with antibody-based therapies as gemtuzumab ozogamicin (GO) with induction and post-consolidation chemotherapeutic regimens did not improve final outcomes for younger patients⁶⁴. With a lack of success and mounting safety concerns, GO was removed from the market in 2010, and is no longer available commercially⁶⁴. Nevertheless, clinical trials are still underway for the discovery of safe applications of GO including maintenance of the remission state following anti-cancer treatment with trans retinoic acid and arsenic trioxide in acute promyelocytic leukemia⁶⁵. Recent positive results with minimized toxicity and safety concerns in the latest GO phase III trial have suggested that there are applications for GO where benefits outweigh risks that manifested in prior applications⁶⁵. Further work in this area has led to second and third generations of ADCs, where components of third generation vadastuxumab talirine include anti-CD33 humanized monoclonal antibody with two drug warheads of pyrrolobenzodiaezpine connected by protease cleavable linker⁵⁶ (Fig. 1). For greater depth on drug discovery in AML, see the review article by Sallman et al.⁶⁶.

Figure 1. Components of antibody-drug conjugates for targeting acute myeloid leukemia.

(A) Structure of first generation antibody-drug conjugate, gemtuzumab ozogamacin, including humanized anti-CD33 IgG4 monoclonal antibody, hydrazine linker at lysine residues, and calicheamicin drug. (B) Components of third generation vadastuxumab talirine including anti- CD33 humanized monoclonal antibody with two drug warheads of pyrrolobenzodiaezpine connected by protease cleavable linker. (C) Components of third generation antibody-drug conjugate IMGN779 including humanized IgG1 monoclonal antibody with three drug units consisting of a cleavable disulfide linker and indolinobenzodiazepine. (D) Summary of benefits and trade-offs for utilizing first, second, or third generation antibody-drug conjugates for treatment of acute myeloid leukemia. Adapted with permission from⁵⁶.

Treatment of CLL has shown success in phase I study of Moxetumomab Pasudotox, an immunoconjugate with specificity for CD22⁶⁷. When compared to rituximab and cladribine – the current standards of care for CLL– fewer side effects were observed, and healthy lymphocyte populations showed better rebounding post-treatment. Despite promising results, several patients treated with Moxetumomab pasudotox were forced to end treatment before complete response after developing antibodies against the immunotoxin⁶⁷. As such, Moxetumomab pasudotox was recently redesigned into a new immunotoxin, LMB11, with less antigenicity⁶⁸. Reduced immunogenicity in LMB11 was achieved by removing excess regions and modifications to surface amino acids of the Pseudomonas exotoxin⁶⁸. Because the designs for reduced immunogenicity were made with humans in mind, mouse studies could not be performed to demonstrate reduced immunogenicity in humans⁶⁸, although antigenicity studies suggest that LMB11 has a longer half-life in the blood while being tolerated at high doses in mice⁶⁸.

Antibody engineering for CML has taken a different approach than engineering for other leukemias. Monoclonal antibodies have been designed to block the interleukin-1 receptor accessory protein (ILRAP/IL1R3) which is vital for IL-1 signaling pathways, and is expressed more consistently than IL-1 in CML⁶⁹. IL-1 signaling has been shown to promote strong proliferation in CML stem cells and is exclusively expressed on primitive CML cells and not healthy hematopoietic stem cells⁶⁹. Specific expression of ILRAP therefore allows for antibody-dependent cellular cytotoxicity and phagocytosis by NK cells and macrophages, respectively⁶⁹.

Several antibody-based therapeutics have also been developed to combat multiple myeloma. In recent research history, monoclonal antibody therapies for multiple myeloma target CD48⁷⁰, signaling lymphocytic activation molecule family 7 protein (SLAMF)⁷¹, and IL-6⁴⁷. CD48 has been proposed as a target form monoclonal antibody therapies because it is known to specifically be expressed by cells within the hematopoietic lineage⁷⁰. This protects the patient from off target binding that is unrelated to their myeloma. CD48 was found to be expressed on over 90% of multiple myeloma cells in 22 patients out of 24 in a pilot study by Hosen et al, proving that it could be a powerful target for antibody therapy⁷⁰. That same research group went on to develop a murine anti-CD48 monoclonal antibody to drive myeloma specific cytotoxicity⁷⁰. Impressive results were shown in a subcutaneous tumor mouse models despite limited evidence of antibody dependent cell cytotoxicity in in vitro studies⁷⁰. Due to the murine origin of the pilot antibody, human trials were unable to be performed to show the full potential of anti-CD48 monoclonal antibodies in treating multiple myeloma, although the potential targeting of healthy lymphocytes and monocytes may cause the risks of the such a therapy to outweigh the benefits in such a trial.

As a result, other attempts at monoclonal antibody therapy have been made for the treatment of multiple myeloma. SLAMF has been proposed as an alternate target to CD48, because it is more widely expressed in myeloma cells while being independent of cytogenic abnormalities inherent to the myeloma⁷¹. The humanized IgG1 monoclonal anti-SLAMF antibody (elotuzumab) was approved by the FDA in 2015 after a successful phase 3 clinical trial in relapsed myeloma patients⁷¹. Although elotuzumab resulted more frequently in adverse reactions of all severities, the clinical trial showed it increased the response rate and extended the period for additional treatment a full year when compared to lenaliodomide plus dexamethasone⁷¹. In addition to CD48 and SLAMF, IL-6 was assessed as a target for antibody therapy for the treatment of myeloma. Slituximab is an chimeric anti-IL6 monoclonal antibody that showed promising applications for myeloma and Castleman’s disease⁴⁷. Despite successful phase one trials, further exploration in a phase two study failed to show that outcomes in relapsed myeloma patients were improved by inclusion of Slituximab with bortezomib when compared to bortexomib in combination with a placebo⁷².

Modest and insignificant success with monoclonal antibodies for myeloma has triggered recent development of antibody-drug conjugates for its treatment. SGN-CD352A, an anti- SLAMF (CD352) cysteine antibody conjugated to pyrrolobenzodiazepine has been shown to interact with myeloma cells through clathrin-dependent endocytosis⁷³. Endocytosis of SGN-CD325A was followed by caspase 3 and 7 dependent apoptosis had minimal affect to off target cell types⁷³. Preclinical in vitro and in vivo studies in mouse xenograft models suggest that SGN-CD325A could prove to be a viable option for treating multiple myeloma in clinical trials⁷³. Antibody-drug conjugates for treating multiple myeloma have also been recently developed targeting B-cell maturation antigen (BCMA) with monomethyl auristatin F (MMAF)(GSK2857916)⁷⁴, CD46 with MMAF⁷⁵, and CD56 with the maytansinoid DM1⁷⁶. While the anti-CD56 ADC, lorvotuzumab mertansine, has moved onto clinical trials (NCT00991562), the specificity towards CD46 (CD46-ADC) is the most novel myeloma antibody-drug conjugate developed recently. This approach is notable because of the previously underappreciated targeting of CD46. Although, CD46 is not a surface marker required, by myeloma cells or other B-cell lineage cells, it has been shown to be upregulated in myeloma cells because of the copy number duplications on chromosome 1q, but remains unexpressed outside placenta and prostate tissues⁷⁵. Chromosome 1q amplifications are a known marker for poor prognosis in multiple myeloma⁷⁵, so targeting this subset of myeloma is a relatively untapped niche of myeloma research and should, therefore, be developed further.

While antibody production and engineering has been investigated for decades now, antibody and human enzyme based-research is still evolving. Some of the most promising results for harnessing the host’s own immune cells have been through the administration of bispecific T-cell engager (BiTE) antibodies. BitE antibodies bind to both the target cell as well as a cytotoxic T cell to drive cell-mediated cytotoxicity. BitE antibodies have been produced with anti-CD19^77–79 for B-cell lineage leukemia and lymphoma cell types paired with anti-CD3^{78, 79} or anti-CD5⁷⁷ regions to join T cells to the cancerous targets. The resulting connection between target cells and the T cell promotes immune synapse formation and target cell lysis⁷⁹. In addition, recombinant human engineered enzymes have also been proposed as a potential therapy in AML. Pegylated human recombinant arginase I cobalt (HuArgI(Co)-PEG5000, AEB1102) was developed to combat arginase accumulation in AML cells⁸⁰. Depletion of arginase stores in AML cells was shown to hinder growth in all nine AML cell lines tested and without off-target effects on healthy hematopoietic cells⁸⁰. Varying levels of arginine dependence were observed, but the dependence was present in all cells tested. Translation of AEB1102 into mice and monkeys proved successful⁸¹, and clinical trials are currently underway to show efficacy in treating relapsed or refractory acute myeloid leukemia as well as myelodysplastic syndrome (clinicaltrials.gov, NCT02732184).

3. Novel small molecule inhibitor targets of signaling molecules and epigenetic mediators in hematological malignancies

The standard combination chemotherapy, CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) was the frontline therapy for DLBCL for >30 years. Rituximab (R-CHOP) was the only major improvement in spite of tens of millions of dollars in expenditure on clinical trials, and still a significant percentage of DLBCL patients are not cured⁸². Therefore, alternative drug treatments are needed to target this particularly difficult malignancy. The initiation and progression of cancer is controlled by both genetic and epigenetic alterations, where the former is almost impossible to reverse. In contrast, epigenetic aberrations are potentially reversible, allowing the malignant cell population to revert to a more normal state. Therefore, in addition to the antibody and emerging protein approaches, identification of signaling pathways and small molecule inhibitors of these targets have emerged as promising highly specific alternatives with low toxicity that potentially reprogram the cancer cells to make them more susceptible to death. Such is the case of enhancer of zeste homologue 2 (EZH2), a histone methyltransferase⁸³epigenetic target, which is required for B cells to form germinal, centers (GCs). Expression of EZH2 is linked to a highly proliferative state of cells not only found in cancer, but in healthy stem cells as well⁸³. Prostate cancer, DLBCL, and breast cancer are just a few types of cancers where EZH2 over expression is a driving forces for tumorigenesis^{83, 84}. Somatic mutation in histone modifying enzymes is the genetic hallmark of GCB-DLBCL. Gain of function mutations of H3K27 methyltransferase EZH2, and loss of function mutations of H3K27 histone acetyltransferases CREBBP, EP300 and H3K4 methyltransferase KMT2D are present in >80% of patients and are early hits during disease pathogenesis^85–92. Recent work has shown that each of these mutations causes lymphomas through aberrant epigenetic programming of their respective histone marks^{85, 87, 93}. Beyond the effects of its mutations, EZH2 is also required for the growth of DLBCLs with wild-type EZH2^{85, 87, 93}. In addition, mutations in the SET domain of EZH2 has been identified in DLBCL^94–96. Since the discovery that EZH2 inhibitors can halt the progression DLBCL, several unique drugs have been under development to specifically reduce EZH2 function including Tazemetostat (EPZ-6438) for NHL^97–99, El1 for DLBCL⁸³, GSK126 for DLBCL⁹⁰, CPI-360 for NHL¹⁰⁰, Compound 3 in EZH2 mutant GCB-DLBCL⁸⁴, and EPZ011989 for DLBCL¹⁰¹. As more cancer types are discovered to be dependent on EZH2 mutations and drug specific resistance emerges in the treatment of those cancers, there will always be a need for the development⁹⁸ of a plethora of EZH2 inhibitors especially in DLBCL¹⁰².

Similar to EZH2 inhibitors, other methylation effectors have been successfully inhibited for targeting DLBCL. DNA methyltransferase (DNMT) 1, DNMT2, and DNMT3 are proteins that methylate at CpG dinucleotides in the DNA sequence and the effector sites of epigenetic regulator drugs¹⁰³ as indicated in Fig. 2. While DNMT2 and DNMT3 have larger role in initiating methylation at new sites, the function of DNMT1 is more closely aligned with the maintenance of preexisting methylation sites. Targeting DNMTs has potential to halt methylation, thereby reversing gene silencing by allowing for initiation of gene transcription¹⁰⁴. 5-aza-2′-deoxycytidine and azacitidine are drugs that inactivate DNMTs by an irreversible formation of a covalent bond between the two upon association with DNA as pyrimidine nucleoside analogs¹⁰⁴. As a result, methylation groups are steadily reduced at cytosine residues in daughter strands of DNA upon cell division effectively turning gene transcription back on¹⁰⁵. Because of the heavy requirement of successive cell division on the successful function of DNMT-targeting drugs, recent efforts to promote their application focused on combination therapy with DLBCL as the target malignancy because of its highly proliferative nature¹⁰⁶. Another way to understand the role of epigenetic modulators is their involvement in immune response. Treatment with DNMT, histone deacetylase (HDAC), and EZH2 inhibitors, lead to enhanced antigen production, antigen processing, CD40 expression, PDL-1 checkpoint target expression in tumor cells, major histocompatibility complex (MHC) class 1 expression, and production of chemokine, cytokines, and interferons^107–109 (Fig. 3A). At the same time, HDAC inhibitors can increase the expression of receptors on natural killer (NK) cells and ligands on AML cells, promoting AML blast killing¹¹⁰. Similarly, HDAC inhibitors have been shown to enhance CD8+ T cell killing and regulation of regulatory T cells in tumors^111–113.

Figure 2. Effector sites of epigenetic regulator drugs.

Upregulation of DNMTs and HDACS alter gene transcription leading to human cancers including hematological malignancies. Similarly, loss of TET supporting proteins leads to decreased gene transcription. DNMT, IDH1, and IDH2 inhibitors inhibit DNA methylation pathways to mitigate loss of function, gene down regulation from methylation. Other epigenetic regulators function through histone modifications that can lead to abnormal chromatin compaction. These include HDACs which reduce acetylation and histone methyl transferases (HMTs) which promote histone methylation. EZH2 inhibitors target the EZH2 HMT to reduce histone methylation to correct gene dysregulation. Adapted with permission from¹⁰³.

Figure 3. Effect of epigenetic inhibitors on immune cells and modes of epigenetic drug delivery.

(A) Epigenetic drugs offer multi-modal approach to targeting malignant B cells. While EZH2 inhibitors exclusively affect malignant B cells, HDAC inhibitors (HDACi) also enhance chemokines, interleukins, and interferons while inhibiting the proliferation of regulatory T cells. Similarly, DNMT inhibitors (DNMTi) enhances nKGG2D receptors on natural killer cells while enhancing MHC class I, PDL-1, tumor antigens, and cytokines from malignant B cells. (B) Targeted drug delivery to cancer cells function through two pathways: cytotoxicity of cancer cells and epigenetic reprogramming through various epigenetic regulating drugs. Nanoparticle drug delivery platforms improve outcomes through enhanced drug stability and sustained drug release over soluble drug delivery paradigms.

In addition to epigenetic reprogramming drugs, recent discoveries have looked into B and T cell signaling pathways as potential target for suppressing tumors. One prominent target is B cell receptor (BCR) pathway. Biological dependence on the BCR signaling pathway is a major hallmark of ABC DLBCLs^{2, 114–119}. The BCR is a transmembrane protein located on the outer surface of healthy and malignant B cells¹²⁰. It is a heterodimer composed of heavy-chain and light-chain immunoglobulins, CD79A/Igα and CD79B/Igβ. BCR phosphorylation facilitates recruitment of proteins including Bruton’s tyrosine kinase (BTK), phospholipase C (PLC)-2, and phosphatidylinositol (PI)-3 kinases. The BCR pathway is chronically activated in ABC-DLBCLs in part due to somatic mutations of CD79A/B (~20% of ABC-DLBCLs)¹¹⁴, CARD11 (~10%)¹²¹, and several others. Given this scenario, a number of BCR pathway targeted therapies are in development – including kinase inhibitors targeting SYK, BTK, PI3K, PKC, MAPKs and AKT¹¹⁷. Among these, ibrutinib, an irreversible BTK inhibitor, has shown activity in clinical trials for ABC-DLBCL patients who have CD79 mutations, although patients with mutations affecting proteins downstream of BTK, such as CARD11 are resistant^{115, 122}. Moreover resistance can develop due to mutations in BTK and other downstream BCR components that are directly or indirectly dependent on tumor microenvironment^{119, 123, 124}.

Another key pathway target for BCR in ABC-DLBCLs is mucosa-associated lymphoid tissue lymphoma translocation protein 1 (MALT1). MALT1 is an attractive target for ABC-DLBCLs and other BCR-dependent lymphoma subtypes. Many of the activating mutations of the BCR pathway cause constitutive MALT1 paracaspase activity, which in turn explains, at least in part how lymphoma cells maintain NFκB activation. Interestingly, translocations of MALT1 lead to: 1) overexpression of the protein or 2) formation of the fusion protein API2-MALT1 in MALT lymphoma^125–127. Amplification of MALT1 locus has been linked to increased MALT1 mRNA expression in ABC-DLBCL¹²⁸. Moreover, a loss-of-function shRNA screening, determined that MALT1 was necessary for ABC but not GCB-DLBCL survival¹²⁹. MALT1 is the only paracaspase in the human genome and hence its enzymatic pocket is structurally unique compared to other caspase family members, offering the possibility of developing highly specific inhibitors. Loss of MALT1 paracaspase activity is less toxic than targeting NFκB signaling more broadly as demonstrated by the fact that MALT1 knockout mice are born at the expected Mendelian rate, are fertile, and present only immune response activation defects^{130, 131, 132}. Accordingly, studies show that ABC-DLBCLs are biologically dependent on and addicted to MALT1^132–136. Hence targeting MALT1 has the potential to impact a broad cross section of ABC-DLBCL patients without causing toxicity to other organs and with reduced likelihood of off-target effects. Along these lines, Melnick and colleagues have reported a small molecule compound MI-2 as a potent and irreversible MALT1 enzymatic inhibitor¹³² while others have reported phenothiazines as reversible allosteric MALT1 inhibitors (e.g. Mepazine, an anti-psychotic drug)¹³⁵.

Similarly, small molecule inhibitors have been developed for the treatment of leukemia. Isocitrate dehydrogenase 1 and 2 (IDH1 and IDH2) inhibitors (Fig. 2) have emerged as powerful treatments for AML. Isocitrate dehydrogenase is an enzyme that is responsible for catalysis of isocitrate to alpha-ketoglutarate¹³⁷. It manifests in two locations with IDH1 in the cytoplasm and IDH2 in the mitochondria of the cell¹³⁷. IDH1 and IDH2 are linked with progression through differentiation and upregulation or mutation of these enzymes are found in some AML patients¹³⁷. Mutations at R172 and R140 of IDH2 are present in approximately 15–20% of AML cases¹³⁷. Inhibition of function has been shown to be a viable treatment for IDH-related cancers, and as a result several small molecule inhibitors have emerged to inhibit IDH1^137–144. While many IDH inhibitory drugs have been discovered over the years, few drugs have gone beyond in vitro validation. The IDH2 inhibitor AG-221 was one of the few to even progress to phase I clinical trials (trial number NCT 01915498 additional also found online at ClinicalTrials.gov)^{137, 138}.

Despite increased numbers of small molecule inhibitors being discovered and engineered for the treatment of hematological malignancies, one major hurdle to their successful application is the method of delivery. Systemic administration of these drugs may result in unwanted side effects or at least inefficient treatment. As a result, drug delivery platforms have been developed to solve this issue by targeting healthy and malignant immune cell populations and effectively usher the use of small molecule inhibitors into modern cancer treatment. Nanoparticle and microparticle engineering has been proposed to solve this problem by acting as drug vehicles for targetted toxicity to cancer cells and sustained drug release in lymphoma^145–147, leukemia^148–150, and myeloma^151–154. Nanoparticle research has built on basic principles of nanomedicine to design silver and gold nanoparticles to specifically target tumors of interest^{149, 150, 154}. For example, a simian immunodeficiency virus accessory protein, Vpx delivered through virus-like particles inhibited the function of deoxynucleoside triphosphate (dNTP) triphosphohydrolase SAM domain and HD domain 1 (SAMHD1), which has been shown to be critical in turning over ara-CTP pools within cancerous cells, and thereby enhanced AML susceptibility to ara-C therapy¹⁵⁵.

While many drugs can have devastating effects on cancer cells, they often come with undesirable side effects and general toxicity resulting from systemic rather than targeted delivery. While some particles engineered for AML actually lack AML cell specificity¹⁵⁰, nanoparticle cancer specificity has been achieved by aptamer modification of gold nanoparticles for the targeted delivery of daunorubicin to the ALL cell line Molt-4¹⁵⁴. In this application, Danesh et al. used Sgc8c, an aptamer specific for tyrosine kinase 7 in T cell leukemias, to drive nanoparticle contact with ALL cell lines¹⁵⁴. Aptamer-cell connections, as used here, deliver chemotherapeutics directly to targeted cells while minimizing off target drug delivery and potentially the undesirable side effects that come with systemic chemotherapy. Similarly, alternative targeting techniques use monoclonal antibodies to drive nanoparticle delivery to hematological malignancies. Significant strides to this field were achieved through Fe₃O₄ nanoparticles with conjugated ACBG2 monoclonal antibody and paclitaxel (McAb-PTX-NPs)¹⁵¹. In the comparative in vitro and in vivo study, the engineered nanoparticles outperformed all other standards of care tested including melphalan and prednisone in combination¹⁵¹. In addition to the McAb-PTX-NPs, simpler drug conjugated nanoparticle systems have also been proposed for treatment of myeloma using the chemotherapeutics carilzomib¹⁵³, and daunorubicin¹⁵⁴.

Epigenetic effector drugs present unique problems with drug delivery because of their quick degradation and excretion rates in vivo. Enzymes like cytidine deaminase in the liver and other tissues have been shown to cause increased degradation of the epigenetic drug decitabine^{156, 157}, thereby minimizing therapeutic effects of treatment. Effectively similar, due to the relatively small size of these drugs and inability to bind to other proteins in the body, excretion and clearance rates are also fast, therefore requiring continuous infusion for therapeutically relevant levels of drug function¹⁵⁸. Nanoscale delivery systems can offer protection from degradation and cancer-targeted delivery (Fig. 3B), and have been a natural solution for epigenetic drug delivery. These delivery methods include dendrimers, nanogels, nanoparticles, and liposomes, and the diversity among them provide tailor made properties for different malignancy types. Liposomal carriers with encapsulated anacardic acid (epigenetic drug) in the liposomal bilayer with a vitamin C gradient, loaded with mitoxantrone has been developed¹⁵⁹. The efficacy of N-isopropylacrylamide (NIPAM)-based biodegradable nanogels with decitabine was shown in drug resistant solid tumors and leukemia cells. Decitabine sustained release from nanogels prolonged cell arrest in the G2/M cell-cycle phase, and enhanced efficacy of the drug¹⁶⁰. Most hematological malignancies allow for increased access for drug delivery, but in the case of solid tumors, nanoparticles and liposomes have been shown to aggregate at tumor sites due to leaky lymphatics and vasculature to overcome general drug inaccessibility^{161, 162}. As more of the epigenetic and signaling targets as well as nanoscale delivery methods are being developed, the diversity of their therapeutic potential in treating the various types of hematological malignancies is increasingly being understood.

4. Targeting the immune cell populations using vaccines and drug delivery systems

For some time now, vaccination has been proposed as a possible tool for the elimination of cancer by the patient’s own immune system. With the proper antigen presentation to cytotoxic T cells, cancer cells can be targeted similarly to old or virally infected cells. Cyclin D1, a frequently upregulated cyclin in mantle cell lymphoma has been singled out as an effective antigen for vaccination therapy¹⁶³. The frequent over expression of cyclin D1 in mantle cell lymphoma and its potency in T cell activation upon cross presentation contribute to why cyclin D1 makes an excellent target for vaccines¹⁶³. Along other lines, Mattarollo et al. showed that a cancer antigen cocktail with an α–galactosylceramide (α-GalCer) adjuvant could successfully activate natural killer T cells against B cell lymphoma and subsequently inhibit its growth in mice lymphoma models¹⁶⁴. Similar strategies have been translated to the targeting of leukemia. An α-GalCer adjuvant vaccine with irradiated acute leukemia cells was tested in mice as a possibility for targeting leukemia¹⁶⁵. Unfortunately, the vaccine proved to only be effective at preventing onset and relapse rather than mitigating an already established leukemia¹⁶⁵. Other approaches utilized a multi-epitope vaccine for CML by delivering the fusion protein BCR-ABL and Wilms’ tumor 1 (WT1) epitopes¹⁶⁶. Lentiviral delivery of these antigens into primary human dendritic cells successfully stimulated autologous CD3+ T cells in vitro¹⁶⁶. While validation of this system has yet to be done in vivo, lentiviral delivered vaccines contain significant promise for the future of leukemia treatment. Cancer vaccines have also been developed for the treatment of myeloma when first line proteasome inhibitors and immunomodulatory agents fail. An mucin 1 (MUC1) targeted vaccine (ImMucin) has been developed for myeloma patients who have not seen responses though traditional treatment pathways¹⁶⁷. MUC1 is a myeloma associated surface marker that can be recognized through a 21-mer peptide loaded vaccine¹⁶⁷. The mechanism of action for the MUC1 vaccine relies on already present MUC1 specific MHC complexes that, after binding, mount a T cell response against the MUC1 expressing myeloma cells¹⁶⁷. Initial phase I/II studies have shown that ImMucin can initially stabilize myeloma levels although there is no indication that it improves patient outcome in the long-term¹⁶⁷.

Recruitment of the adaptive immune response through antigen recognition, processing, and presentation followed by T cell activation is critical for cancer vaccine function. Vaccine efficiency has been known to be hindered due to early antigen degradation prior to the recognition stage. Both DNA- and protein-based vaccines are susceptible to both extracellular and intracellular degradation, the latter which prevents antigen translocation in the cytoplasm or to the nucleus. Encapsulating drug delivery platforms such as micro- or nanoparticles offer protective properties from in vivo degradation, and as such, have been proposed as cancer vaccine carriers^168–170. Polyesters^{171, 172}, polyanhydrides¹⁷³, and liposomes¹⁷⁴ have served as effective base biomaterials for micro- and nanoparticles in vaccine delivery. Upon delivery to sites of malignancy, immunostimulatory and immunomodulatory components such as pathogen-associated molecular patterns are used to target immune cells such as dendritic, tumor, or T cells. For example, cationic microparticles with polyethyleneimine attached to the surface, improve cell transfection, and activation of APC by the idiotype pDNA antigen (A20 idiotype-chemokine (MCP-3) fusion plasmid), yet they survive harsh endosomal environments^{175, 176}. Altogether, this vaccine platform led to anti-tumor immunity in a prophylactic model of B cell lymphoma importantly without adjuvant support.

Another major strategy in particle engineering, is the utilization of small interfering RNAs (siRNAs)^{147, 148, 152}, CpG oligodeoxynucleotides (CpG ODN)¹⁴⁵, or both¹⁴⁷. siRNA block expression of target genes in tumor cells. Both oncogenes and signaling cytokines have been targeted using the technique to reduce proliferation and alter the immune response, respectively^{147, 152}. CpG ODN on the other hand, act in an adjuvant fashion to drive maturation of dendritic cells though binding to toll-like receptor 9 to allow for tumor antigen presentation to cytotoxic T cells and an increased Th1 immune response¹⁴⁵.

In addition to micro- and nanoparticle delivery systems, multi-modal biomaterial delivery systems have expanded on the importance of porosity, degradability, and general physical structure for vaccine and drug delivery^177–179. For example, dendritic cell migration through delivery materials for antigen discovery was promoted by increasing the size of material pores, faster in vivo degradation rates, cytokines, and danger signal presentation^{178, 179}. All the while, this system allowed dendritic cells to migrate out of the material to transition to draining lymph nodes and initiate the immune response. This feat was achieved by an injectable in situ cross linking dextran vinyl sulfone and tetra-functional polyethylene-glycol thiol hydrogel that degrades quickly in vivo for effective delivery of plasmid DNA antigen, siRNA for inhibition of IL-10, and MIP-3α for dendritic cell recruitment^{178, 179}. This work considers several factors that affect multi-modal delivery systems including matrix gelation time (which can be affected by reactive group substitution), the effects of physiological temperature, the dependence of swelling ratio on polymer combinations, and the promotion of dendritic cell invasion by material class and porosity^{178, 179}. Ultimately, all of these factors came together to allow for a multimodal fast-degrading in situ crosslinking hydrogel system that can effectively shift immune response towards Th1 responses, despite poor immunogenicity of B cell lymphoma in a mouse model, leading to notably increased survivorship when challenged with a tumor cells in lethal quantities^{147, 179}.

Since siRNAs require intracellular access to properly perform their function of blocking gene expression, additional approaches have been developed using lipid nanoparticles that fuse with the cell membrane for transport into the cell¹⁴⁸. This drug delivery platform, SLP-301R, has been shown to prefer transfection with suspension leukemia cell lines with low expression of the Cav1 and Cav2, cell membrane surface lipid rafts, that have historically be difficult to transfect with standard lipid nanoparticle systems¹⁴⁸. The transfection interaction was correlated to the decreased expression of the Cav1 and Cav2 genes, that are also under-expressed in neuronal cells, which are also difficult to transfect¹⁴⁸. This may suggest that while siRNA delivery was only presented in the bone marrow and spleen, there could be potential off target siRNA delivery in the nervous system, and that further research into lipid nanoparticles should address misdirected drug delivery to healthy tissues.

5. Cell Engineering Approaches to Target B and T cell Tumors

In recent years, the engineering of immune cells has emerged to the forefront of hematological cancer treatment. Through lentiviral delivery, chimerical antigen receptors (CAR) are introduced into natural killer and T cells for a cancer cell targeted cytotoxic cellular response. First generation CARs link the antigen receptor of choice to a CD3ζ domain for intracellular signal transduction, while second generation CARs also include either CD28 or 4–1BB (CD137) as costimulatory membrane proteins^{180, 181}. Engineered CAR cells have been implemented in both T cells^{180, 182–187}, natural killer cells (NK)^{188, 189}, and cytokine induced killer cells (CIK)^{190, 191} to channel their cytotoxic functions at CAR-complemented cancer cells. They have been developed for AML, CLL, multiple myeloma, B cell lymphoma, and ALL. AML targeting CAR cells typically express anti-CD33^{182, 191} or anti-123^{184, 190} CARs. While both showed success in targeting AML blasts, comparison of these two target surface proteins showed that targeting CD123 had less off target cytotoxicity to hematopoietic progenitor cells than anti-CD33 CAR T CIKs^{190, 191}. Engineered CAR cells for AML all used primary human cell lines either collected through apheresis from peripheral blood draws or from umbilical cord blood. Emphasis was placed on autologous cell sources for T cells to maintain HLA restriction for in vivo applications (NCT02623582 NCT01864902 NCT02799680).

Major efforts towards fighting hematological malignancies have used CD19 targeted CAR T cells (NCT03029338, NCT02822326, NCT02975687, NCT02842138, NCT02546739, NCT02656147, NCT02081937). Anti-CD19 CAR T cells specifically target healthy and malignant B cells, and have implemented for CLL¹⁹², B cell ALL^{183, 186, 193}, and follicular lymphoma¹⁹⁴. In all applications, autologous T cells were used from with mixed ratios of CD4 and CD8 positive T cells that were highly patient dependent¹⁸⁷. Despite the differences in cytotoxicity profile of the harvested patient T cells, no correlation was found between the ratio of cell type and patient outcome¹⁸⁷. Patient outcomes were improved though through at least 5 rounds of oral ibrutinib (a Bruton’s tyrosine kinase inhibitor of the BCR pathways) prior to anti-CD19 CAR T cell infusions¹⁹². Ibrutinib not only targeted the malignant B cells, but reduced expression of programmed cell death 1 protein, an immunosuppressive membrane protein in T cells¹⁹². In addition to attacking malignant cells, engineered CAR cells have been proposed to tackle other issues with cancer treatment. In B cell ALL, Epstein Bar Virus- and cytomegalovirus-specific autologous memory T cells transfected with anti-CD19 CARs have been suggested to protect immunosuppressed patients from these viruses after hematopoietic stem cell transplant, and through specific effector T cell selection, reduce the risk of graft versus host disease triggered by the CAR T cell therapy¹⁸³. Anti-CD19 CAR T cells have also been applied to follicular lymphoma^{185, 194}. Recent research in this area has shown that by programming CAR T cells to also produce soluble herpes virus entry mediator (HVEM), the effectiveness of engineered cell therapy can be enhanced¹⁹⁴. This could provide the framework for future directions in engineered CAR T cells through the expression of soluble effector molecules to enhance cancer targeted cytotoxicity.

Multiple myeloma targeted CAR therapies have also been developed targeting CD138 (syndecan 1)¹⁸⁹, B-cell maturation antigen (BCMA)^{195, 196}, and CS1^{180, 188}. CD138, BCMA, and CS1 are highly expressed on multiple myeloma cells but are more importantly lowly expressed on hematopoietic stem cells^{180, 189}. CS1 has proved a successful target, as shown by the success of the CS1 targeted therapy Elotuzumab¹⁸⁸. While engineering of CS1 cells relied on primary T¹⁸⁰ and NK¹⁸⁸ cells, Jiang et al. took a different approach and used the NK-92MI cell line, a derivative of large granular lymphoma cells¹⁸⁹. Cell lines can be used in this case because NK cells are not MHC/HLA restricted and cell lines offer greater ease for in vitro expansion prior to therapy. To mitigate risks associated with cell therapy using immortalized cell lines, the anti-CD138 CAR NK cells were irradiated by 10 Gy to stop cell division without a reduction in cytotoxic effect when tested in a myeloma xenograft mouse model¹⁸⁹. Even with these interesting outcomes with CD138 targeted therapy, CS1 has been shown to be more widely and specifically expressed in patient myeloma cells than CD40, CD56, CD138, and CD74¹⁸⁸. Although CS1 may be a better target for CAR T cell therapy conceptually, clinical trials for anti-CD138 CAR T cells (NCT01886976) and anti-BCMA CAR T cells are currently underway¹⁹⁶ (NCT02658929, NCT02215967, NCT02954445, NCT02786511).

6. Engineered ex vivo models of hematological malignancies and their role in drug discovery

While many cancer studies begin with experiments on cancer cell lines, they all must be studied in an animal model before clinical translation. Mouse models tend to be the animals of choice because of their small size, relative cost-effectiveness, and mammalian similarities to humans. All hematological malignancies reviewed here possess mouse models to mimic disease within the complexity of a complete organism. Many animal models for hematological malignancies begin with NOD/Scid mice to which cancerous cell lines or patient explants can be introduced and allowed to flourish. Introduction of cancer cell lines intravenously, intramuscularly, or subcutaneously has been the standard platform for hematological malignancy mouse models for some time. Examples of this technique are found in AML^{140, 141, 155, 184, 191}, ALL¹⁹⁷, DLBCL^{101, 164, 198–200}, and multiple myeloma^{180, 188, 201}. For liquid tumors, this alone is sufficient for a working cancer model, but in palpable tumors like lymphoma, cells are often injected with 50% Matrigel as the scaffold for lymphoma cells to inhabit^{97, 200}. These models have been the standard for preclinical research because they recapitulate spontaneous tumorigenesis in human cancers and allow for the use of human cancer cell lines over mouse cell lines that may not accurately represent human disease. Immortalized human cancer cell lines though, have their drawbacks. Through countless passages in vitro, cell phenotype can be altered. Older, more established cell lines then may behave differently when reintroduced in vivo than newer cells. In fact, when gene expression of 90 cell lines were compared to primary tumor cell samples, the cell lines tended to cluster with greater similarity to other cells lines regardless of tumor origin than the primary cancer cells²⁰². In addition, through the cell culturing process, established cell lines have indirectly been selecting for rapidly dividing cells over cells with slower proliferation²⁰². This selection process reduces cell line heterogeneity and could account for discrepancies in cell line based and clinical trials.

For this reason, patient derived xenografts (PDX) have been proposed to better recapitulate human cancers. In a PDX, patient malignant cells or tumorous tissue are engrafted into immunodeficient mice as a model for the patient’s specific cancer. This allows for exact matching of oncogenic mutations and minimizes culture time ex vivo that can lead to genotypic and phenotypic changes in cancer cell lines. Mouse PDX models provide the patient cells with organismal context to better recapitulate the systematic complexities of cancer. Solid patient tumors are often passaged three or more times between animals, and this process has shown to alter the integrity of the tumor stroma²⁰³. While typical patient tumor explants contain human tumor and stroma, through prolonged implantation, stromal cells, matrix components, nutrient supplying vessels drift to overwhelmingly murine in origin²⁰³.

Within the many rotating parts of xenograft models of, it can be difficult to discern which elements of the xenograft are eliciting the observed response in the context of pharmacological studies. Because many of the elements of xenografts are out of the researcher’s control, three dimensional in vitro models of hematological malignancies have been developed for AML^{204, 205}, follicular lymphoma^{206, 207}, B cell lymphoma^{208, 209}, and multiple myeloma^210–214. With the modularity of tissue engineering, these 3D models are highly adaptable to be disease specific. While culturing techniques like using the hanging drop method for the formation of aggregates^{207, 215} required little engineering for microenvironment development, microenvironment tailoring has been achieved through manipulation of biomaterial stiffness²⁰⁹, strength²¹², integrin ligand presentation^{208, 209}, inclusion of stromal cell support^{204, 213}, and morphology²⁰⁵. Even though tailored tissue engineered constructs are over-simplifications of the tumor microenvironment, the presence of these 3D structures in cancer drug trials has been shown to produce different responses when compared to traditional 2D cell culture^{204, 206–209, 212}. For example, our group has developed specific integrin ligand presenting designer 4-arm polyethylene glycol hydrogel organoids of B and T cell lymphomas²⁰⁸. The biological findings in the study showed heterogeneity in the expression of various integrin receptors across subclasses of B cell lymphomas and T cell lymphomas. The presentation of integrin ligand-specific amino acid polymer segments, in a designer and tailored fashion, provide a biological connection between cancer cells and their supporting extracellular structure. Transmission of these survival cues through integrins, has been shown to alter lymphoma cell phenotype within the context of chemotherapy resistance and promotion of overall survival^{208, 216, 217}. While the mechanisms of the promotion of survival vary between cancer types and some remain unclear, B cell lymphomas survival was correlated to the presence of stromal support cells and the upregulation of IgM response the 3D engineered microenvironment^{208, 217}. As these mechanisms become more fully understood, it is becoming increasingly clear that the extracellular environment, including biochemical and mechanical cues, significantly alters cancer cell response to chemotherapy in vitro^{218, 219}. As such, greater effort should be made to improve in vitro models of B and T cell tumors in the hopes to surpass animal models in their predictive power while improving statistical power from high throughput capabilities all the while reducing study costs for drug discovery^{218, 220}.

7. Conclusion and Perspective

As cancers continue to evolve in the presence of chemotherapies, there will likely always be a need for new anti-cancer drugs in the treatment of the majority of hematological malignancies. As such, efforts should remain not only in drug discovery, but in discovery platforms as well. A simple schematic of future drug research is summarized in Figure 4. While many of the reviewed methods have shown efficacy in targeting hematological malignancies, multi-faceted or novel approaches have come to be seen as the future of these fields of cancer research. One example of such a therapeutic would be the anti-CD20 nanodisks developed by Crosby and colleagues. This approach utilizes apolipoprotein A-I nanoparticle chemotherapy delivery systems that are typically protected by a protein ring, and replaced typical ring structures with a single chain variable antibody fragment specific to CD20 for B-cell malignancy targeting²²¹. The resulting fusion protein, termed nanodisks, were shown to specifically deliver curcumin, an anti-oxidant polyphenol, to CD20 positive lymphoma cells in vitro with successful intracellular release upon binding²²¹. The result of this process was decreased viability in Ramos, a CD20 positive B cell lymphoma cell, further showing the effectiveness of this nanodisk drug delivery system and the promising therapeutic potential it holds. Similar lipid nanoparticle targeting techniques have also been developed for multiple myeloma by conjugating very-late antigen 4-specific ligand to lipid nanoparticles loaded with carfilzomib¹⁵³. Additional novel approaches to targeted hematological malignancy therapy include the gene therapy drug delivery platform engineered by Dong et al.²²². In this approach, lentiviral delivery of suicide genes was directed by an anti-MUC1 antibody for gene therapy targeted to leukemia cells²²². The suicide genes included herpes simplex virus thymidine kinase and Escherichia coli cytosine deaminase followed by treatment with the prodrugs ganciclovir and 5-fluorocytosine, respectively. Lentiviral delivery of suicide genes allowed for intra-leukemia cellular drug activation leading to leukemia cell death by direct and bystander mechanisms that are yet to be fully understood²²². This leukemia therapy represents a targeted therapy by the anti-MUC1 antibody that minimizes off-target cytotoxicity. In addition, they targeted leukemia cells with two drugs for a multifaceted approach and greater leukemia cell death. Moving forward, smart and intelligent drug delivery system that could function in an “on-demand” fashion is warranted to tackle relapsed tumors. The authors imagine these systems to circulate in the body for very long period and prevent any relapse tumor buildup. Finally, we proposed the translation of cancer drug discovery and initial drug trials from two dimensional cultures to tissue engineered constructs at least in the case of solid hematological tumors. While they may not provide identical scenarios to human disease, they continue to provide more patient representative features of growing tumors when compared to traditional cancer monocultures. For drug discovery to shift away from traditional cell cultures, greater efforts should be made to develop tissue engineered constructs that are simple, replicable, scalable, and representative of both the healthy and diseased tissues they aim to recapitulate. In essence, many such new platforms for drug discovery and drug delivery will have to consider the genetic and epigenetic heterogeneity of hematological malignancies, which continues to remain a major roadblock in the field.

Figure 4. Proposed pathway for drug discovery and translation of therapeutics in hematological malignancies.

Drug discovery is often required on a special patient- and mutation-specific basis. Understanding the oncogenic profile of patients can be achieved through genomic, metabolic, and proteomic analyses. With greater understanding of the mutation types from the patient, tumor specific cytotoxicity can be evaluated in genetic mouse models, patient derived xenograft models, or in ex vivo engineered tissues, each with distinct advantages and disadvantages. Each of these models can be used to evaluate efficacy and toxicity of drugs, followed by formulation development using biomaterials with reduced toxicity effects. Drug delivery platforms could be engineered for both single-mode and combinatorial delivery of drugs to establish improved treatment protocols for patient disease. Upon discovery of tumor-targeting drug and suitable delivery platform, dose response studies evaluating anti-tumor effect, genomic, metabolic, and proteomic changes in tumor cell response can initiate new clinical trials to treat human hematological malignancies.

Glossary of Acronyms and Abbreviations

ABC

Activated B cell-like

ADC

Antibody-drug conjugates

ALL

Acute lymphoid leukemia

AML

Acute myeloid leukemia

BCMA

B-cell maturation antigen

BiTE

Bispecific T-cell engager antibodies

BTK

Bruton’s tyrosine kinase

CAR

Chimerical antigen receptor

CCR4

CC motif chemokine receptor 4

CHOP

Cyclophosphamide, doxorubicin, vincristine, and prednisone

CIK

Cytokine inducer killer cells

CLL

Chronic lymphocytic leukemia

CML

Chronic myeloid leukemia

CpG ODN

CpG oligodeoxynucleotides

DLBCL

Diffuse large B cell lymphoma

DNMT

DNA methyltransferase

dNTP

Deoxynucleoside triphosphate

EZH2

Enhancer of zeste homologue 2

FcR

Fc receptors

GC

Germinal center

GCB

Germinal center

GO

Gemtuzumab ozogamacin

HDAC

Histone deacetylase

HVEM

Herpes Virus Entry Mediator

IDH1

Isocitrate dehydrogenase 1

IDH2

Isocitrate dehydrogenase 2

IgG

Immunoglobulin G

IL

Interleukin

ILRAP/IL1R3

Interleukin-1 receptor accessory protein

MALT1

Mucosa-associated lymphoid tissue lymphoma translocation protein 1

MAX

MYC-associated factor X

MHC

Major histocompatibility complex

MMAF

Monomethyl auristatin F

MUC1

Mucin 1

MXD3

MYC-associated factor X dimerization protein 3

NIPAM

N-isopropylacrylamide

NK

Natural killer

PD-1

Programmed cell death protein 1

PD-L1

Programmed cell death ligand 1

PDX

Patient derived xenografts

PI

Phosphatidylinositol

PLC

Phospholipase C

R-CHOP

Rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone

SAMHD1

Triphosphohydrolase SAM domain and HD domain 1

siRNA

Small interfering RNA

SLAMF

Signaling lymphocytic activation molecule family 7 protein

WT1

Wilms’ tumor 1

α-GalCer

α–galactosylceramide