Fifth-week immunogenicity and safety of anti-SARS-CoV-2 BNT162b2 vaccine in patients with multiple myeloma and myeloproliferative malignancies on active treatment: preliminary data from a single institution

It is highly desirable that vaccines against SARS-CoV-2, hopefully including its variants, could work also in hematological malignancy patients. Such patients were not enrolled in the registration trials, and, so far, no data on COVID-19 vaccines for this vulnerable population have been published, except for chronic lymphocytic leukemia [1]. Therefore, recommendations and expectations in clinical practice are based on evidences coming from the immune response to COVID-19 infection and non-COVID-19 vaccines [2].

In a large retrospective multicenter Italian cohort study, overall COVID-19-attributable mortality rate in 536 patients with hematological malignancies was 37% [3]. Patients with progressive disease status and diagnosis of acute myeloid leukemia, lymphomas and plasma cell neoplasms had worse outcomes. The mortality was 37% among the 106 patients with plasma cell neoplasm. A similar overall COVID-19 mortality rate of 34% was also observed in a large retrospective International Myeloma Society Study on 650 patients with multiple myeloma (MM) from 10 different countries [4]. Multivariate analysis showed that age, high-risk disease by FISH, renal disease and suboptimal control of myeloma were independent predictors of worse outcome. Lower mortality rates were observed in patients with myeloproliferative neoplasms (MPN) (myelofibrosis, polycythemia vera, essential thrombocytosis). In a large European observational retrospective study, 50 of the 175 patients (29%) with MPN died at a median of 9 days after the diagnosis of COVID-19 [5]. Mortality was higher than in the general population and reached 48% in patients with myelofibrosis. There are also clinical and laboratory observations on a protective role of tyrosine kinase inhibitors (TKIs) versus SARS-CoV-2 infection in patients with chronic myeloid leukemia (CML) [6, 7].

COVID-19 vaccines recommendations in hematological patients have been provided by several scientific societies [8‐14]. In summary, mRNA vaccines should be preferred, and while there are no particular expected problems in terms of safety, the efficacy data have to be verified in light of potentially reduced immunogenicity in immunocompromised patients. These patients are on the one hand at higher risk of COVID-19 mortality, on the other hand at higher risk of ineffectiveness from vaccination, both aspects due to deep humoral and cell-mediated immunosuppression related to treatments and underlying disease. A real uncertainty exists about the optimal timing, dosing and capacity of some subgroups of hematological patients to efficiently respond to the COVID-19 vaccination [15, 16]. Several guidelines in pre-COVID-19 era recommended that vaccines are administrated at least 3–6 months after the end of an anti-lymphoid therapy in order to avoid their likely futility [17, 18]. Two consensus-based recommendations in MM patients were very recently published by the European Myeloma Network and an experts Italian panel [19, 20]. Both publications highlighted that the response in myeloma patients is often less vigorous.

On the basis of the Italian national plan against COVID-19, a RNA vaccine should be offered as soon as possible to onco-hematological patients in treatment with immunosuppressive or myelosuppressive drugs or within 6 months from the end of such treatment, and to stem cell transplanted patients after 3 months from transplant [21].

According to this plan, since March 2021 BNT162b2 mRNA vaccination (Pfizer-BioNTech) has started at our institute for patients with hematological malignancies on active treatment.

Herein we report preliminary data on 42 patients with MM and 50 with myeloproliferative malignancies (MPM) (Philadelphia-negative MPN n = 30 and CML n = 20), who were vaccinated and evaluated for anti-SARS-CoV-2-neutralizing IgG titer on day of first injection, on day of second injection (i.e., after 3 weeks) and after 2 weeks from second injection, according to a monocentric institutional study. Serological responses were compared with those observed in 36 elderly subjects aged over eighty not suffering from cancer.

Two mRNA vaccines have been approved by FDA and EMA, having demonstrated to enhance effectively immunogenicity against SARS-CoV-2 in general population [24, 25]. In particular, BNT162b2 was 95% effective in preventing COVID-19 [25]. Hematological patients were not enrolled in the registration trials. The extent of a hypothetical reduced efficacy of mRNA vaccines to elicit an immunological response respect to general population, although expected in these immunocompromised patients, is widely unknown.

Four issues regarding vaccination in hematological patients become dramatically crucial in a pandemic like that by COVID-19.

Firstly, defining the positivity cutoff of reliable tests that measure neutralizing antibodies is the methodological premise for any subsequent reasoning. LIAISON® SARS-CoV-2 S1/S2 IgG test by DiaSorin® was validated as a marketable CLIA by FDA. The positivity cutoff for neutralizing IgG was defined using data from wild infection in COVID-19 individuals. Are there consistent immune-biological or technical reasons why it might not be valid even after vaccination or in immunocompromised patients? An antibody titer evaluated with LIAISON® SARS-CoV-2 S1/S2 IgG above 15 AU/mL has demonstrated a concordance level with a gold standard for the evaluation of the virus-specific antibody neutralizing power represented by PRNT90 at 1:40 ratio of 94.4%, while above 80 AU/mL the level of agreement with 1:160 PRNT90 titer was 100%. There are no valid reasons in our opinion to doubt that this commercial CLIA is reliable in detecting the neutralizing antibody concentration at the cutoff of 15 AU/mL even after vaccination and in immunocompromised patients.

Secondly, defining the optimal timing of vaccination is a paramount challenge. There is no doubt that vaccinating patients at least one months before starting treatment or 3–6 months after the end of therapies would be optimal. Unfortunately, there are cancer patients who cannot stop ongoing therapies or who cannot wait. There is a need to take responsibility what to do for such vulnerable patients, particularly those receiving lympho-depleting treatments. Our policy was to vaccinate all patients in active treatment according to indication of the national plan as soon as possible, without substantial modifications or delay in the administration of the ongoing treatment scheme. Therefore, it should be noted that all the data from our study were obtained in patients on active treatment for their hematological neoplasm.

Thirdly, in patients with hematological malignancies immune deficiencies are heterogeneous. Patients with lymphoproliferative disorders and those with myeloid neoplasms suffer from distinct immune deficiencies and receive different treatments that variously affect the vaccine response and that are notably different compared with the past. Our choice was to focus on two different set of diagnoses, poles apart in many ways from each other.

Fourthly, although the incidence of proven SARS-CoV-2 infection in the vaccinated subjects remains the preferred clinical efficacy endpoint of a vaccine, the seroprotection rate has to be regarded as an acceptable surrogate endpoint in immunocompromised populations. Therefore, it was important to immediately verify the immunogenicity of a mRNA vaccine in hematological patients and to compare the data with those obtained from a non-neoplastic elderly control population similar aged.

At the best of our knowledge, this is the first report focusing on the humoral response to mRNA BNT162b2 vaccination in patients with MM and MPM.

As proof of concept, the mRNA BNT162b2 vaccine has demonstrated to be immunogenic in immunocompromised patients with MM and MPM on active treatment, eliciting a significant serological response. The GMCs of IgG anti-S1/S2 subunits of SARS-CoV-2 spike protein significantly rose after the booster in all the cohorts, at a significantly different extent among the cohorts. In fact, the highest absolute increase of IgG after the booster, expressed in terms of GMC, was reached in the elderly controls, the lowest in the MM patients, with MPM patients in the middle.

MPM patients responded to vaccine reaching proportions of seroconversion of 88%, significantly lower compared to controls over 80 aged (100%; p = 0.038). MM patients responded partially, with a seroconversion rate of 78.6% after the booster significantly inferior to elderly controls (p = 0.003). Therefore, a near-complete seroprotection was achieved in MPM patients, whereas the booster worked at minor extent in patients with MM. Furthermore, such seroprotection in MM patients does not appear robust, since almost one-third of MM patients, defined responders at the cutoff of 15 AU/mL, would be defined not responders by raising the cutoff at 80; in fact, at such cutoff, seroprotection rate decreases from 78.6 to 54.8%, contrary to what happens in the other two cohorts.

These observations, although exciting, are not surprising.

MM is one of the most immunologically defective conditions among cancers per se. A recent review summarized dysfunctions of immune cells in the myeloma niche [26]. From cross-talk between tumor plasma cells and bone marrow (BM) microenvironment derives an immunosuppressive context characterized by loss of effective antigen presentation, dysfunction of effector cells, expansion of immunosuppressive cells. As a consequence, the immunogenic effectiveness of vaccines in MM is a hot issue. In patients with MM, a yearly single dose of inactivated influenza vaccine is recommended by ECIL 7 guidelines in light of controversial benefit of a booster, and variable seroconversion rates, mostly between 20 and 25%, have been observed [27]. However, at least in two studies dedicated to MM patients, which measured the rates of seroprotection after each of two doses of influenza vaccine, the rates of patients with protective titers rose significantly post-booster, reaching proportion around 70% [28, 29]. Active disease requiring therapy was associated with lower likelihood for a serological response. Also in a randomized placebo-controlled trial of adjuvanted recombinant zoster vaccine (two doses 1–2 months apart) administered during or after immunosuppressive cancer treatment in 569 patients with hematological malignancies, including 132 MM patients, the humoral response at month 2 was 80%. Unfortunately, separate results for MM patients were not reported [30]. These percentages are not different from that observed in our study. It should be extremely relevant to establish whether there are identifiable biomarkers predicting the vaccine response. From this point of view, treatment was significantly associated with response to BNT162b2 vaccine. Patients on active treatment with proteasome inhibitors-based and imids-based therapies, alone or in combo, without daratumumab, had a significant likelihood of responding to vaccine (92.9%). On the contrary, daratumumab in combo with lenalidomide was significantly associated with a lower response rate (50%). This is in line with the current evidences. In fact, even if treatments on the one hand restore a condition of partial immune-competence by direct anti-neoplastic effects and disruption of interaction between plasma cells and BM microenvironment, on the other hand, selectively impacts on many cellular actors of the immune system [26]. At one end of the spectrum, imids promote immune activation by functional enhancement of T and NK cells, increase of Th1 cytokine production, reduction of Treg activity, improvement of dendritic cells maturation and functions, enhancement of antibody-dependent cell-mediated cytotoxity [31‐33]. At the other end, daratumumab (and isatuximab) targets CD38 on the population of normal and tumor plasma cells, in this way reducing vaccines immunogenicity by direct depletion of antibody producer cells. However, a recent paper reported that IgG levels and induction of protective antibody titers were intact against Streptococcus Pneumoniae, Hemophilus Influenzae B and seasonal influenza at a median of two months after treatment, probably by escape from daratumumab of a subset of plasma cells due to a reduced expression of CD38 [34]. The policy of not modifying the therapy scheme before, during and after the vaccine administration suggests cautiousness in the interpretation of our data, which might be even better outside an active therapy setting. However, the fact that not all MM patients responded to vaccine, and that those who responded did it not robustly, poses great concerns and turns to the opportunities offered by the new mRNA vaccination platforms. For example, repeated booster of mRNA vaccine, at intervals to be defined, could theoretically reinforce the humoral response. Clinical trials should determine the optimal timing and dosing schedule of BNT162b2 vaccination in this patient group. Presently, it appears clear that increasing measures are urgently required to consolidate the scarce immunological protection achieved by vaccination in these MM patients. For the purposes of a correct risk management, it appears important to highlight both the need to vaccinate the family context, and the need that no misleading message should be conveyed through the mass media on the complete effectiveness to prevent COVID-19 by vaccination of such population, who still has to be protected by virtuous behaviors based on the use of social distancing, face mask wearing and hand hygiene.

The low percentages of SARS-CoV-2 infection and severe COVID-19 which were observed in patients with CML and chemo-free-treated Philadelphia-positive acute lymphoblastic leukemia might be related to a protective action exerted by TKIs [6, 35]. It has been recently reported that genes with anti-infective effect (for example CCL5, CD28, IFN gamma) are upregulated and others with pro-infective effect (for example ARC-1, FUT4) are downregulated by TKIs in patients with CML [7]. The expression profiles of these genes involved in inflammation and immunity were measured after six months of treatment with imatinib, and the results supported the idea of an absence of immunological damage after a median-length treatment with TKIs. From our data, all the 20 patients with CML on TKIs responded robustly to BNT162b2. In Philadelphia-negative MPN patients, the COVID-19 mortality was higher than in general population, particularly in myelofibrosis, and was correlated with the interruption of ruxolitinib, probably due to enhancement of cytokine release syndrome which in turn may lead to multiorgan failure [5]. No conclusion on how ruxolitinib works on vaccine immunogenicity may be drawn from our population because only six patients were on ruxolitinib at the time of vaccination. Also the number of patients with myelofibrosis (n = 8) was too small to be analyzed as variable in its own right; however, 4 out of these 8 patients did not respond to vaccine. Overall, data from other vaccination contexts in patients with CML or Philadelphia-negative MPN are limited to inactivated influenza vaccine and pneumococcal polysaccharide 23-valent vaccine and to patients not receiving ruxolitinib [27]. In particular, in patients with CML, a strategy based on two doses of H1N1 vaccine administered three weeks apart allowed to increase the seroprotection rate from 85 to 95%, proportions not so different by those observed after vaccination with BNT162b2 in our cohort of 20 CML patients [36].

Our study has some limitations: First of all, the limited number of patients impacts on the statistical power when we look at some variables as predictors of response, although some trends seem clear (i.e., myelofibrosis as variable potentially associated with a lower likelihood of response). Moreover, the limited observation period and the absence of concomitant investigations on the T-cell response do not allow to draw firm conclusions on two paramount questions which remain unanswered, namely whether responding patients actually develop true protection from COVID-19 and how long this protection extends over time. Combining cellular and humoral measures of vaccine efficacy may increase the ability to predict the risk of COVID-19. Evidence exists of robust memory T-cell responses in patients with MPN following SARS-CoV-2 infection [37]. The extent and robustness of such T-cell response after vaccination with BNT162b2 is a crucial point to verify in both MM and MPM patients. Recently, delayed and impaired B- and T-memory cells responses were reported at three weeks and three months after a single dose of seasonal influenza A vaccine in MPN compared to healthy controls [38].

At five weeks, the population with MPM responded robustly and near completely to BNT162b2 vaccine with the exception of patients with myelofibrosis, closing the gap from elderly controls. In MM patients, who responded partially and less robustly, anti-CD38-based therapy impacted negatively on the vaccination. The new mRNA platforms allow to hypothesize different administration schedules based on closer and/or personalized booster, opening the doors also in this scenario to a precision medicine. In light of the scarce immunological protection by vaccine, in these MM patients on active treatment with anti-CD38 monoclonal antibodies, IgG titer monitoring should be implemented and social distancing and mask wearing maintained regardless of vaccination status. Primary prophylaxis by anti-SARS-CoV-2 monoclonal antibodies should be explored in trials for MM patients at higher risk of no response. However, confirmatory and more exhaustive studies are urgently needed.